Balance Between Senescence and Apoptosis Is Regulated by Telomere Damage–Induced Association Between P16 and Caspase-3

JBC Papers in Press. Published on May 10, 2018 as Manuscript RA118.003506 The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.RA118.003506

Balance between senescence and apoptosis is regulated by telomere damage–induced association between p16 and caspase-3

Shanmugam Panneer Selvam1,2, Braden M. Roth1,2, Rose Nganga1,2, Jisun Kim1,2, Marion A. Cooley3,#, Kristi Helke4, Charles D. Smith5, and Besim Ogretmen1,2*

1Department of Biochemistry and Molecular Biology, 2Hollings Cancer Center, 3Department of Regenerative Medicine, 4Department of Comparative Medicine, Medical University of South Carolina, 86 Jonathan Lucas Street, Charleston, SC 29425; 5Department of Pharmacology, Pennsylvania State University, 500 University Drive, Hershey, PA 17033.

#Current Address: Department of Oral Biology, Augusta University, Augusta, GA 30912. *Address correspondence to: [email protected]

Running title: p16-caspase-3 complex and senescence Key words: apoptosis; senescence; sphingolipid; sphingosine 1-phosphate; telomere damage

Downloaded from Abstract damage–induced apoptosis, indicating that an Telomerase activation protects cells from telomere association between p16 and caspase-3 damage by delaying senescence and inducing cell proteinsforces senescence induction by inhibiting

immortalization, whereas telomerase inhibition caspase- 3 activation and apoptosis. These results http://www.jbc.org/ mediates rapid senescence or apoptosis. However, suggest that p16 plays a direct role in telomere the cellular mechanisms that determine telomere damage–dependent senescence by limiting damage–dependent senescence versus apoptosis apoptosis via binding to caspase-3, revealing a induction are largely unknown. Here, we direct link between telomere damage–dependent

demonstrate that telomerase instability mediated senescence and apoptosis with regards to aging by guest on May 31, 2018 by silencing of sphingosine kinase 2 (SPHK2) and and cancer. sphingosine 1-phosphate (S1P), which binds and stabilizes telomerase, induces telomere damage– Introduction dependent caspase-3 activation and apoptosis, but Telomerase is composed of a catalytic reverse not senescence, in p16-deficient lung cancer cells transcriptase (TERT) and telomeric template RNA or tumors. These outcomes were prevented by (TR), which plays a key role in the elongation and knockdown of a tumor-suppressor protein, maintenance of telomeres at chromosome ends (1- transcription factor 21 (TCF21), or by ectopic 4). Shelterin complex proteins associate with expression of WT human telomerase reverse telomere repeats for protection from inducing transcriptase (hTERT), but not mutant hTERT DNA damage response to overcome replicative with altered S1P binding. Interestingly, SphK2- end protection problem (5-8). Telomerase is highy deficient mice exhibited accelerated aging and expressed in stem cells, while it is inactivated due telomerase instability that increased telomere to epigenetic silencing of TERT upon damage and senescence via p16 activation differentiation in most somatic cell types (9,10). especially in testes tissues, but not in apoptosis. However, telomerase is reactivated in majority of Moreover, p16 silencing in SphK2-/- mouse cancer cells, inducing cell immortalization and embryonic fibroblasts activated caspase-3 and tumorigenesis (11-14). Recent studies revealed a apoptosis without inducing senescence. Further, clinical relevance of TERT activation in cancer, as ectopic WT p16 expression in p16-deficient A549 mutations within the core promoter region of the lung cancer cells prevented TCF21 and caspase-3 TERT gene provide binding sites for E-twenty-six activation, and resulted in senescence in response (ETS) transcription factors, leading to cancer- to SphK2/S1P inhibition and telomere damage. specific telomerase reactivation (15-17). Mechanistically, a p16 mutant with impaired Activation of telomerase protects cells caspase-3 association did not prevent telomere from telomere damage, delaying senescence and aging process (18,19), whereas targeting (Fig. 1A-B). Similarly, inhibition of SPHK2 using telomerase induces tumor suppressor responses ABC294640 (100 mg/kg, 7 days/week treatment such as senescence or apoptosis (20,21). However, for 21 days) suppressed A549-xenograft-tumor signaling mechanisms that distinctly induce growth compared to vehicle-treated controls (Fig. senescence versus apoptosis in non-cancerous 1A). Interestingly, ABC294640 treatment had no versus cancerous cells in response to telomerase additional effect on A549-xenograft-tumor growth inactivation and telomere damage induction are suppression in response to shRNA-mediated largely unknown. SPHK2 knockdown, supporting its role in SPHK2 Sphingosine 1-phosphate (S1P) is a inhibition for tumor suppression (Fig. 1A). bioactive sphingolipid, generated by sphingosine Suppression of A549-xenograft-tumor growth in kinase 1 or 2 (SPHK1 or SPHK2), that exerts pro- response to SPHK2 knockdown or inhibition was oncogenic signaling via engaging with its G- consistent with decreased hTERT protein protein coupled receptors (S1PR1-5) in an abundance and apoptosis measured by IHC using autocrine or paracrine manner, and/or via directly anti-hTERT antibody and TUNEL assay, binding to regulate its intracellular target proteins respectively, compared to control tumor tissues without receptor signaling (22-25). Intracellular (Fig. 1C-D). In addition, SPHK2 inhibition or targets of SPHK2-generated nuclear S1P include knockdown resulted in increased telomere damage HDAC1/2 (26) and TERT (27). Our previous in A549-xenograft-tumors compared to controls, Downloaded from study described a role for SPHK2-generated S1P measured ex vivo by immunofluorescence using in the stabilization of telomerase via directly anti-TRF2 and anti-g-H2AX antibodies in TIF binding to hTERT involving its D684 residue (27). (telomere damage-induced foci) assay (31) (Fig. These studies also suggested that S1P binding 1E). Moreover, there was no significant difference http://www.jbc.org/ mimicked the phosphorylation of hTERT at Ser921, in the staining of senescence associated-beta which protected hTERT from ubiquitination and galactosidase in shSPHK2 tumors compared to proteasomal degradation (27). However, roles and shScr controls (Fig. 1F). Similar data were also mechanisms by which targeting SPHK2/S1P and obtained when A549-luciferase derived tumors TERT induces senescence versus apoptosis with were grown orthotopically in the lungs of SCID by guest on May 31, 2018 respect to telomere damage signaling and tumor mice after tail vein injection, and inhibition of suppression are largely unknown. Thus, in this SPHK2 using ABC294640 decreased lung tumor study, we have outlined experiments to uncover growth (~95%) and hTERT abundance compared signaling mechanisms that determine responses for to vehicle-treated controls (Supplemental Fig. the induction of senescence versus apoptosis via S1B-D). telomere damage-mediated cellular stress. To assess the clinical relevance of hTERT regulation by SPHK2/S1P signaling, we then Results determined SPHK2 abundance with regard to Inhibition of SPHK2/S1P attenuates hTERT hTERT expression by IHC using commercially abundance and tumor growth available tumor microarrays (TMAs) containing To determine the roles of SPHK2/S1P in lung tumor tissues obtained from patients with the regulation of tumor suppression via controlling NSCLC (n=48) and their adjacent pathologically telomerase and telomere damage, we measured the non-cancerous lung tissues (n=48). The data effects of silencing and/or pharmacologic showed that SPHK2 is significantly over- inhibition of SPHK2/S1P signaling using shRNA- expressed (~2.5-fold) in the majority of NSCLC dependent knockdown or small molecule inhibitor tumors compared to their controls (46/48 tumors, of SPHK2, ABC294640 (28-30) (Supplemental p<0.05) (Supplemental Fig. S2A-B). The data also Fig. S1A) on the growth of A549 xenograft- showed that hTERT was up-regulated (~2.8-fold) derived tumors in SCID mice. ShRNA-mediated in the majority of NSCLC tumors (47/48, p<0.05) knockdown of SPHK2 (~75% mRNA reduction (Supplemental Fig. S2B). Importantly, the compared to Scr-shRNA-expressing tumors, Spearman correlation analysis showed that there is measured by qPCR) almost completely prevented a significant (n=48, p<0.001) association between A549-xenograft-tumor growth compared to Scr- SPHK2 and hTERT expression in these tissues shRNA/A549-xenograft controls after 21 days (Supplemental Fig. S2C). Similar correlation

2 between SPHK2 and hTERT was detected in the activation in response to ABC294640, catalytically majority of tumor tissues obtained from patients inactive mutant SPHK2G212E expression (34) was with small cell lung cancer (SCLC) (Supplemental not protective compared to controls (Fig. 2D-E). Fig. S2D). Thus, these data suggest that targeting Exogenous addition of S1P (5µM, 1h) also clinically relevant and upregulated SPHK2/S1P in prevented ABC294640-mediated caspase 3 tumors inhibits hTERT abundance, which is activation in A549 cells compared to vehicle- associated with increased telomere damage and treated controls (Fig. 2F). apoptosis, leading to tumor growth suppression in Increased telomere damage in response to vivo. genetic loss of SphK2 was also detected in SphK2- /- compared to WT-MEFs (Fig. 3A). Moreover, Targeting SPHK2/S1P signaling mediates shRNA-mediated SPHK2 knockdown resulted in telomere damage and subsequent caspase 3 telomere damage in A549 cells compared to Scr- activation in lung cancer and/or immortalized shRNA-transfected controls (Fig. 3B). As a MEFs control, we also detected the effects of hTERT Because SPHK2 silencing or inhibition knockdown on telomere damage in A549 cells in induced apoptosis (increased TUNEL staining) in the absence/presence of ABC294640. ShRNA- ABC294640-derived lung tumor xenografts in mediated hTERT knockdown and SPHK2 SCID mice (Fig. 1D), we then measured the inhibition using ABC294640 induced telomere Downloaded from effects of ABC294640 on caspase activation using damage at a similar level (~2.5-fold) compared to MEFs obtained from caspase3/caspase7 knockout controls (Fig. 3C). However, inhibition of SPHK2 mice (32), containing their various combinations had no additional effect on telomere damage in of homozygous or heterozygous deletions. response to hTERT knockdown (Fig. 3C). Treatment of caspase3+/-/caspase7+/- MEFs with Moreover, shRNA-mediated knockdown of http://www.jbc.org/ ABC294640 (80 µM, 8 h) increased caspase 3/7 SphK2 or its pharmacologic inhibition using activity ~2-fold compared to vehicle treated ABC294640 resulted in telomere damage and controls (Fig. 2A). Deletion of caspase 3 alone or subsequent caspase 3 activation in A549 cells in in combination with caspase 7 completely response to G0/G1 arrest during serum starvation by guest on May 31, 2018 abolished caspase activation in response to (Supplemental Fig. S3A-B and S3C-D). In ABC294640 in caspase3-/-/caspase7+/- or caspase3- addition, shRNA-mediated knockdown of SphK2 /-/caspase7-/- MEFs compared to controls (Fig. 2A). (Supplemental Fig. S4A) decreased S1P and Interestingly, deletion of caspase 7 in caspase3+/- dihydro-S1P without causing large changes in /caspase7-/- MEFs increased basal levels of caspase C16-ceramide or dihydro-C16-ceramide compared 3 activity compared to vehicle-treated caspase3+/- to Scr-shRNA-transfected cells (Supplemental Fig. /caspase7+/- control MEFs, and ABC294640 was S4B). These data suggest that telomere damage able to further induce caspase activity in and/or caspase 3 activation in response to caspase3+/-/caspase7-/- MEFs compared to their SphK2/S1P knockdown is regulated independently vehicle-treated controls. Moreover, ABC294640 of cell cycle status and/or changes in ceramide and increased telomere damage in a similar manner in dihydro-ceramide accumulation in A549 cells. caspase3+/-/caspase7+/- and caspase3-/-/caspase7+/- However, SphK2 inhibition using MEFs (Fig. 2B), suggesting that caspase 3 ABC294640 mediated telomere damage without activation is regulated downstream of telomere detectable caspase 3 activation in primary non damage in response to SPHK2 inhibition. These cancereous human lung fibroblasts (generation 2, data were also consistent in A549 cells, in which Supplemental Fig. S5), suggesting that caspase 3 ABC294640 induced caspase 3 activity ~1.8-fold activation downstream of telomere damage might compared to vehicle-treated controls, and be regulated differentially in A549 lung cancer inhibition of caspase 3 using Z-DEVD (33) versus non-cancereous lung fibroblasts. pretreatment completely abrogated ABC294640- Interestingly, inhibition of SPHK2/S1P signaling mediated caspase 3 activation (Fig. 2C). In using ABC294640 had no effect on p53-induced reconstitution experiments in A549 cells stably DNA damage as measured by co-localization of expressing Scr or SPHK2 shRNAs, while ectopic p53 and g-H2AX using immunofluorescence in expression of WT-SPHK2 protected caspase 3 A549 cells (Fig. 3D). Cisplatin exposure was used

3 as a positive control for p53-mediated DNA death in response to ABC294640 compared to damage, which increased the co-localization of controls (Fig. 5C). However, phosphomimic p53 and g-H2AX in A549 cells compared to S921D/D684A-hTERT mutant expression also vehicle-treated controls (Fig. 3D). These data prevented ABC294640-mediated cell death (Fig. suggest that targeting SphK2/S1P signaling results 5C). Thus, these data suggest that S1P-binding to in caspase 3 activation due to selective telomere hTERT involving D684 protects telomere damage damage, and not due to the activation of general and caspase activation by mimicking DNA damage response. phosphorylation of hTERT at Ser921. These data also suggest that targeting SPHK2/S1P signaling SphK2/S1P inhibition leads to telomere damage by ABC29640 induces telomere damage due to through attenuation of S1P-hTERT association attenuation of S1P-hTERT association, leading to and phosphomimicking of telomerase inhibition of phosphomimicking of hTERT, To determine the involvement of hTERT in the hTERT instability, and subsequent caspase 3 regulation of telomere damage and subsequent activation. caspase 3 activation, we performed studies in GM847 cells, which do not express endogenous Activation of tumor suppressor TCF21 TERT, but contain telomeric template RNA (27). mediates telomere damage and caspase 3 In GM847 cells, ABC294640 increased caspase 3 activation in response to SphK2/S1P targeting Downloaded from activity ~3-fold compared to controls, and ectopic To determine the signaling between expression of WT-hTERT abrogated this SPHK2/S1P/hTERT targeting and caspase 3 ABC294640-mediated caspase activation (Fig. activation, we measured the effects of SPHK2 4A-B). However, ectopic expression of a mutant knockdown on the expression of genes involved in http://www.jbc.org/ hTERT with D684A conversion, with decreased lung tumor growth/suppression by q-PCR-based S1P binding, was not protective for caspase 3 lung cancer specific RT2 profiler PCR array (35) activation in response to ABC294640 (Fig. 4A-B). using total mRNA isolated from A549-xenograft- Moreover, expression of the phospho-mimic derived tumors expressing SPHK2 shRNA mutant of hTERT (S921D) with D684A compared to Scr-shRNA expressing tumor by guest on May 31, 2018 conversion, which does not bind S1P controls (shown in Fig. 1A). The data showed that (Supplemental Fig. S3C), was able to protect down-regulation of SPHK2 induced the expression ABC294640-mediated caspase activation (Fig. of GPM6A, CA4, CDH13, TCF21, 4A-B). Expression of WT- and mutant hTERT CDKN2A/OPCML and ERBB2 (~21-3-fold), proteins was confirmed using Western blotting while decreasing the expression of CXCL13, (Fig. 4D). MMP9 and CLCA2 (~4-8-fold) compared to These data were also consistent with the controls (Supplemental Fig. S6A). Among these, effects of ectopic expression of WT-hTERT, but we were able to validate increased expression of not D684A-hTERT with altered S1P binding, on tumor suppressor TCF21 (36) (~5-fold) in the protection of telomere damage in response to response to SPHK2 knockdown in A549 cells SPHK2/S1P targeting using ABC294640 (Supplemental Fig. S6B). Moreover, shRNA- compared to vector-transfected controls (Fig. 5A- mediated knockdown of hTERT also induced B). However, phospho-mimic mutant of hTERT TCF21 mRNA expression (~3.5-fold) compared to restored telomere damage in response to Scr-shRNA-expressing controls (Supplemental ABC294640 even in the absence of S1P binding Fig. S6C). Ectopic expression of WT-hTERT, but due to D684A conversion in cells that express not D684A-hTERT, prevented TCF21 mRNA D684A/S921D double mutation compared to induction in response to SPHK2 knockdown controls, expressing S921A or S921D single (Supplemental Fig. S6D). In addition, targeting mutations, which still bind S1P and protected SPHK2/S1P using ABC294640 induced TCF21 telomere damage (Fig. 5A-B). Similarly, while mRNA (~10-fold) and resulted in caspase 3 ectopic expression of WT-hTERT prevented activation (~8-fold) compared to vehicle treated ABC294640-mediated cell death, measured by controls (Supplemental Fig. S4E-F). ShRNA- trypan blue exclusion assay, inhibition of S1P- mediated knockdown of TCF21 (Supplemental hTERT binding by D684A mutation restored cell Fig. S6E) prevented caspase 3 activation in

4 response to ABC294640 in A549 cells galactosidase expression in their paws (37) at (Supplemental Fig. S6F). generation 6, and decreased weight of testes at To determine the effects of TCF21 generation 4 or 5 compared to WT and SphK1-/- activation in the regulation of lung tumor mice (generation matched controls) (Fig. 7A-B). suppression in response to SPHK2/S1P/hTERT Examining of spleen, skin and testes tissues at targeting in vivo, we measured the growth of generation 6 using H&E staining also confirmed A549-derived xenografts expressing Scr- or increased senescence and aging phenotype in TCF21-shRNAs in the absence/presence of SphK2-/- compared to WT mice. ABC294640. Treatment with ABC294640 (100 SphK2-/- animals at generation 6 exhibited mg/kg at every 7 days for 21 days) inhibited the splenic atrophy with decreased white blood cells growth of Scr-shRNA expressing tumors with no germinal centers present. It should be compared to vehicle treated controls (Fig. 6A). noted that the splenic atrophy is common in aged ShRNA-mediated knockdown of TCF21 prevented mice and can be accompanied with cachexia. ABC29464-mediated tumor suppression compared There was decreased erythrocyte (red blood cell) to controls (Fig. 6A). After growth measurements, turn over in the SphK2-/- mice, which might be due remaining tumors were extracted from these mice, to decreased numbers of erythrocytes in and effects of ABC294640 in the presence/absence circulation. Moreover, we observed decreased of TCF21 knockdown on hTERT mRNA, caspase thickness of the hypodermal adipose layer in the Downloaded from 3 activation and apoptosis were measured. SphK2-/- mice with multiple areas of ABC294640 reduced hTERT mRNA abundance in inflammation, increased fibrosis and abnormal the presence/absence of TCF21 knockdown (Fig. follicle development. Also, there were fewer 6B), suggesting that TCF21 induction is regulated seminerferous tubules (hypoplasia) in SphK2-/- downstream of hTERT down-regulation in mice compared to wild type control testes tissues http://www.jbc.org/ response to targeting SPHK2/S1P signaling. (Fig. 7C). Moreover, ABC294640 resulted in caspase 3 In testes tissues, there was decreased activation (measured by IHC) and apoptosis mTERT abundance (measured by IHC) in SphK2-/- (measured by TEM), which were prevented in compared to WT and SphK1-/- mice (generation 6), by guest on May 31, 2018 response to shRNA-mediated knockdown of whereas there was no apoptosis (measured by TCF21 (Fig. 6C-D). ShRNA-mediated knockdown TUNEL staining) or TCF21 induction (measured of TCF21 was confirmed in A549 xenograft- by IHC) in these mice (Fig. 7D). In addition, derived tumors ex vivo by qPCR (Fig. 6E). Sh- telomere damage induction was detected in the RNA-mediated knockdown of TCF21 had no testes of SphK2-/-, but not in WT or SphK1-/- mice detectable effects on telomere damage in response (Fig. 7E). Thus, these data indicate that loss of to ABC294640-mediated SPHK2 inhibition SphK2 results in accelerated senescence and aging compared to control tumors (Fig. 6F-G). Thus, starting at generation 4, which is consistent with these data suggest a role for TCF21 induction in decreased mTERT and induction of telomere caspase 3 activation and apoptosis in response to damage (especially in testes tissues). However, the inhibition of SPHK2/S1P/hTERT axis and accelerated senescence/aging in mice appeared to telomere damage by ABC294640, leading to be independent of TCF21 and/or apoptosis tumor suppression. induction. To further examine the effects of SphK2 Genetic loss of SphK2 results in accelerated loss, confirmed by Western blotting using anti- senescence and aging via telomere damage and SphK2 antibody (38), in the induction of p16 induction without apoptosis in mice senescence, we have isolated MEFs at generations To determine the physiological impact of loss of 4-6 from WT, SphK1-/- and SphK2-/- mice, and SphK2/S1P signaling on senescence and aging due measured senescence by measuring b-gal and p16 to TERT alterations in a non-cancerous model, we expression. Induction of senescence was detected have generated subsequent generations of WT and in the generation 4, 5 and 6 (G4-G6) of SphK2-/- SphK2-/- mice (for 6 generations). SphK2-/- mice compared to WT and SphK1-/- MEFs Fig. 8A-B). exhibited accelerated senescence and aging In addition, increased telomere damage was phenotype measured by increased beta- detected in SphK2-/- compared to WT and SphK1-/-

5 MEFs (Fig. 8C). Interestingly, while shRNA- SPHK2/S1P/TERT targeting and telomere damage mediated knockdown of p16 had no effect on in lung cancer cells. telomere damage induction in SphK2-/- MEFs, it induced caspase 3 activation compared to controls Interaction between p16 and caspase 3 prevents (Fig. 8D). Collectively, these data suggest that loss caspase 3 activation and apoptosis in response of SphK2/S1P results in accelerated aging and to telomere damage senescence with p16 induction in SphK2-/- To define the mechanism by which p16 activation compared to WT and SphK1-/- mice at subsequent forces cell to undergo senescence but not generations 4-6. apoptosis, we set out studies to test a novel hypothesis that interaction between p16 and Induction of p16 distinctly mediates telomere- caspase 3 might limit caspase 3 activation and damage-mediated senescence and prevents prevent apoptosis. This hypothesis was first tested apoptosis in response to SPHK2/S1P/telomerase in H1341 human lung cancer cells, which express targeting p16. SPHK2 inhibition by ABC294640 mediated To determine whether alterations of the associaton between endogenous p16 and SPHK2/S1P/hTERT signaling induces senescence caspase 3 in H1341 cells, and this interaction was in A549 lung cancer cells, which lack p16 attenuated in response to shRNA-mediated expression (39), we measured the effects of down- silencing of p16 compatred to controls (Fig. 10A). Downloaded from regulation or inhibition of SPHK2 on caspase 3 Similar data were also observed when SPHK2 activation and b-gal induction in the expression was knocked down using shRNA, absence/presence of ectopic expression of p16. which resulted in increased association between ShRNA-mediated knockdown of SPHK2 induced ectopically expressed WT-p16 and caspase 3 in caspase 3 and ectopic expression of p16 attenuated A549 cells with/without ABC294640 exposure http://www.jbc.org/ this process compared to Scr-shRNA expressing compared to Scr-shRNA/vector-transfected and A549 cells (Fig. 9A). ShRNA-dependent vehicle-treated controls using PLA and knockdown or ABC294640-mediated inhibition immunofluorescence (Fig. 10B-C). These data of SPHK2 had no detectable effect on the were also consistent when we measured the by guest on May 31, 2018 induction of senescence in A549 cells, whereas binding between recombinant and purified human ectopic expression of p16 with/without SPHK2 p16 and caspase 3 proteins (40) using OpenSPR targeting increased senescence in A549 cells localized surface plasmon resonance (LSPR) compared to Scr-shRNA-expressing or vehicle- biosensor at various protein concentrations. The treated controls (Fig. 9B). Interestingly, while data showed that purified p16 interacted with ectopic expression of p16 had no effect on the caspase 3 (KD, equilibrium dissociation constant= induction telomere damage (Fig. 9C), it 1.82 x 10-8 +/- 6.5 x 10-9 M) (Supplemental Fig. completely prevented caspase 3 induction in S7A-B). Thus, these data suggest that p16 binding response to SPHK2/S1P inhibition using might prevent the activation of pro-caspase 3. ABC294640 compared to vehicle-treated controls Next, using a molecular simulation and (Fig. 9D-E). In reciprocal studies, shRNA- docking, we generated a model in silico for dependent knockdown of p16 in H1341 human lug protein-protein interaction between p16 and cancer cells (confirmed by qPCR, Fig. 9F) resulted capsase 3 using their exisiting structures in enhanced caspase 3 activation in response to determined by x-ray crystallography (Fig. 10D). SPHK2 inhibition by ABC294640 compared to The model suggested that the interaction between Scr-shRNA transfected controls (Fig. 9G). Lack of p16 and caspase 3 might involve Ser152 of p16 protein abundance of p16 in A549 and H157 (turquoise) and Gly251 of caspase 3 (grey), which human lung cancer cells compared to HeLa and is within the activation domain of caspase 3 H1341 cells, which express p16 is also confirmed (His237-Cys285). To validate this model, we by Western blotting (Fig. 9H). Thus, these data ectopically expressed WT-p16 and mutant-p16 suggest that p16 expression determines the with Ser152 to Ala conversion, and measured their induction of senescence while inhibiting TCF21- association in p16-null A549 cells using anti-p16 dependent caspase 3 induction in response to and anti-caspase 3 antibodies by PLA compared to vector-transfected controls. The data showed that

6 WT-p16 and caspase 3 association was induced in targeting SPHK2/S1P induces hTERT response to SPHK2 inhibition using ABC294640, degradation, resulting in telomere damage and whereas there was no detectable association apoptosis, expression of phosphomimic mutant of between Ser152Ala-p16 in A549 cells (Fig. 10E). hTERT with S921D conversion even in the Moreover, ectopic expression of WT-p16 and not absence of S1P binding due to D684A mutation Ser152Ala-p16, largely protected cells from caspase prevented telomere damage in response to SPHK2 3 activation in response to ABC294640 exposure inhibition. It will be interesting to define whether compared to vector-transfected controls (Fig 10F). phosphomimic function of S1P upon binding to Ectopic expression of WT-p16 and Ser152Ala-p16 other proteins, such as HDAC1/2(26) and in A549 cells was confirmed using Western prohibitin 2 (41), or TRAF2 (22) and PPAR- blotting compared to vector-transfected controls gamma (42), play any roles in the regulation of using anti-p16 and anti-actin antibodies (Fig. their biological functions. 10G). Importantly, interaction between p16 and Interestingly, our current data also caspase 3 was also detected in testes tissues suggested that TCF21-dependent caspase 3 isolated from Sphk2-deficient mice at generation activation plays a key role in tumor suppression 5, which expressed SA-ß-Gal, but not in testes downstream of telomere damage in response to isolated from matched WT or SphK1-/- mice, SPHK2/S1P targeting. TCF21 was reported as a which were also negative for SA-ß-Gal expression tumor suppressor, whose expression is suppressed Downloaded from (Fig. 10H). Thus, these data suggest that by epigenetic methylation in lung tumor tissues interaction between p16 and caspase 3 involves and cells (43,44). It was shown recently that the S152 residue of p16, and that this interaction is TCF21 is reactivated by the expression of long key to prevent caspase 3 activation, which forces noncoding RNA TARID, which enhanced its senescence while limiting apoptotic cell death in demethylation via GADD45A (45). It is, however, http://www.jbc.org/ response to SPHK2/S1P inhibition, and remains unknown whether SPHK2/S1P/TERT subsequent telomere damage. inhibition results in TCF21 activation through TARID/GADD45A. Discussion In these studies, targeting SPHK2/S1P by guest on May 31, 2018 Here, our data revealed that SPHK2-generated signaling using molecular (shRNA) or S1P binds and stabilizes TERT involving its D684 pharmacologic (ABC294640) tools resulted in residue via mimicking TERT phosphorylation at consistent telomere damage in various cell types. S921, which protects telomere damage, and results Although ABC294640 is known to have off target in delayed senescence and protection from effects, especially inducing the accumulation of apoptosis. These data also suggested that dihydro-ceramides due to inhibition of DES targeting/inhibition of the SPHK2/S1P/telomerase (46,47), knocking down of SphK2 using shRNA axis results in telomere damage, which signals had no effect on dihydro-ceramide accumulation TCF21-dependent caspase 3 activation in cancer (Supplemental Fig S4B). Thus, these data suggest cells or immortalized MEFs, or p16-dependent that telomere damage evoked by SphK2 silencing senescence and accelerated aging phenotype in or inhibition appears to be independent of non-cancerous lung fibroblasts, primary MEFs ceramide/dihydro-ceramide alterations. However, (non-immortalized), and non-cancerous testes effects of shRNA-mediated knockdown of SphK2 tissues in subsequent generations of SphK2-/- mice or its inhibition using ABC294640 on S1P versus (Fig. 11). Mechanistically, this study demonstrated ceramide and/or dihydroceramide accumulation in that interaction between p16 and caspase 3, nuclear membranes need to be determined. involving Ser152 of p16, prevents caspase 3 It was interesting that SphK2-/- mice activation in response to telomere damage, forcing showed accelerated telomere damage/senescence cells to selectively induce senescence but not and aging phenotype after four generations. This apoptosis. Our previous study revealed that S1P- was somewhat expected as mTERT-/- showed TERT binding, involving the D684 residue, signs of telomere damage and aging phenotype stabilizes TERT via phosphomimic function of after generation 4 also (48). Interestingly, while S1P at Ser921 to regulate hTERT stability (27). genetic loss of SphK2 induced senescence and Here, the data further suggested that while aging due to telomerase instability and increased

7 telomere damage without induction of apoptosis in Biologics) and 1% penicillin and streptomycin SphK2-/- compared to WT and SphK1-/- knockout (Cellgro). H1299 (expressing wild type K-Ras, mice, downregulation and/or inhibition of and lack expression of p53 due to a homozygous SPHK2/S1P/TERT signaling resulted in TCF21- gene deletion) and H1341 cells were grown in dependent caspase 3 activation down stream of RPMI-1640 (American type culture collection) telomere damage in lung cancer cells. Our data with 10% fetal bovine serum and 1% penicillin further suggested that activation of p16 was the and streptomycin. GM00847 cells were purchased key to decide the fate of cells to succumb in from Coriell Cell Repositories. Wild type, SphK1- senescence rather than apoptosis, as ectopic /-, SphK2-/- MEFs were obtained from K. Argraves expression of p16 in p16-deficient cancer cells (Medical University of South Carolina). GM00847 induced senescence by preventing caspase 3 cells and MEFs were cultured in DMEM and o activation. These data are consistent with a role of incubated at 37 C with 5% CO2. Primary human p16 in the protection of ATR4-mediated DNA lung fibroblasts were purchased from Lonza, Inc. damage in response to dysfunctional telomeres and cultured in fibroblast growth media as (49). Our data are also in agreement that in described (27). addition to p53 signaling, p16 is another effector protein for the telomere damage signaling (50). Plasmids and shRNAs Moreover, recent studies suggest that p16- Plasmid containing hTERTWT in pcDNA vector Downloaded from expressing cells shorten healthy life span (51), and was obtained from J. Chen (Arizona State that targeting p16 to clear senescent cells delays University) and site directed mutagenesis was aging-related disorders detected in adipose tissue, performed to generate hTERTD684A, hTERTS921A, skeletal muscle and eye (52), or rejuvenates aged hTERTS921A-D684A, hTERTS921D, hTERTS921D-D684A hematopoietic stem cells in mice (53). These using Q5 site directed mutagenesis kit (E0554S, http://www.jbc.org/ studies are consistent with our data, which provide New England Bio labs) using manufacturer’s a mechanistic link between p16 and inhibition of instructions. SPHK2WT and SPHK2G212E plasmids caspase 3 and apoptosis, supporting that targeting were obtained from Dr. Sarah Spiegel (Virginia p16 induces clearance of senescent cells possibly Commonwealth University). pLenti-CMV p16- by guest on May 31, 2018 by releaving caspase 3 from p16 for activation of Neo(W111-1) (Plasmid# 22260) and PLenti-CMV caspase 3 and apoptosis. To this end, our data Neo DEST (705-1) (Plasmid#17392) were might have important clinical implications such obtained from Addgene. For knockdown studies, that inhibition of p16 would switch senescent cells were plated at 40-50% confluency and phenotype to apoptosis in tumors, improving Dharmafect transfection reagent I was used as per tumor elimination and suppression, at least in part, manufacturer’s instructions. For SPHK2 in response to SPHK2/telomerase targeting. knockdown, cells were transfected with 200nM Nevertheless, these data provide a novel siRNA (L-004831, Dharmacon, USA) for 72 h and link between the regulation of senescence versus for hTERT knockdown 100nM siRNA (L-003547, caspase 3-dependent apoptosis, which is controlled Dharmacon) was used. by p16 signaling at least in response to telomere SPHK2 shRNA (5’-3’): DNA damage evoked by SPHK2/S1P/TERT CCGGCTACTTCTGCATCTACACCTACTCGA inhibition. GTAGGTGTAGATGCAGAAGT-AGTTTTTG (TRCN00000036973); Materials and Methods Cell lines and culture conditions shTCF21#1(5’-3’): Cell lines were tested for mycoplasma CCGGCGTTTCCAAACCAGAGGAGATCTCG contamination, and authenticated by short tandem AGATCTCCTCTGGTTTGGAAACGTTTTT repeat (STR) profiling (PowerPlex16HS) at the (TRCN0000015563); Genetica DNA Laboratories (Burlington, NC) in November 2016. A549 cells (expressing K-ras shTCF21#2 (5’-3’): mutant, wild type p53) were grown in Dulbecco’s CCGGCGACAAATACGAGAACGGGTACTCG Modified Eagle’s Medium (DMEM) (Cellgro) AGTACCCGTTCTCGTATTTGTCGTTTTT with 10% fetal bovine serum (FBS) (Atlanta (TRCN0000015563); and non-targeting shRNA

8 containing plasmids were purchased from Open synthesizing complementary DNA(cDNA) using Biosystems Inc. Plat-A or Plat-E cells were the iScript cDNA synthesis kit (Bio-Rad). Taqman transfected using the viral transduction protocol as probes (hTERT, Hs00162669_m1; SPHK2, described by the RNA interference Hs00219999_m1; TCF21, Hs00162646_m1; consortium(RNAi). The viral supernatants were Tcf21, Mm00448961_m1; CDKN2A, added to either A549 cells or MEFs, and selection Hs00923894_m1; Cdkn2a, Mm00494449_m1); was performed using puromycin (1µg/ml) for 14 Actb, Mm00607939_s1; RPLPO, days. The following shRNA’s (Sigma) were Hs99999902_m1 were used for quantitative PCR obtained from MUSC shRNA shared technology (Life technologies). resource : sh-p16 (mouse) (5’-3’): Antibodies CCGGGTGAACATGTTGTTGAGGCTACTCGA V5 (R960-25, Invitrogen), SPHK2 (1709-1AP, GTAGCCTCAACAACATGTTCACTTTTTG Proteintech and ab37977, Abcam), g-H2AX (TRCN0000077815); (ab2893, Abcam), TRF-2 (IMG-124A, Imgenex), ki67 (ab15580, Abcam), p53 (#554294, BD sh-p16(Human) (5’-3’): Pharmingen), mTERT (sc-7212, Santa Cruz), CCGGCCGATTGAAAGAACCAGAGAGCTCG hTERT (ab32020, Abcam), POD1(mouse TCF21) AGCTCTCTGGTTCTTTCAATCGGTTTTTG (sc-7294, Santa Cruz), p16 (A301-267A, Bethyl Downloaded from (TRCN0000265833) laboratories) , cleaved caspase-3 (2305-PC-020, Trevigen) antibodies were used. Visualization of S1P-hTERT or p16-caspase 3 interactions using proximity ligation assay Immunohistochemistry and tumor micro array http://www.jbc.org/ (PLA) (TMA) GM00847 cells overexpressing hTERTS921D, Lung adenocarcinoma TMA (LC1002) and Small hTERTS921D-D684A plasmids were fixed with 4% cell lung cancer TMA (LC10010) was purchased formalin for 15 minutes. Fixed and permeabilized from US Biomax and immunohistochemistry was cells were incubated with S1P specific antibody performed with SPHK2 antibody (1:100) and by guest on May 31, 2018 (Sphingomab, LT1002; 20µg/ml) and hTERT hTERT antibody (1:50). Also, primary lung tumor antibody (ab32020, Abcam), or anti-p16 (A301- tissues were obtained from the MUSC tumor 267A, Bethyl laboratories) and anti-caspase 3 biorepository core facility and (2305-PC-020, Trevigen) antibodies for 18 h. PLA immunohistochemistry was performed by fixing was performed and visualized by IF-CM using with formalin, embedding with paraffin and 5µm Duolink in situ hybridization kit as described by sections were made for immunohistochemistry. the manufacturer (Olink Biosciences). Anti-IgG Pathology analysis and scoring was performed by antibody was used as a negative control in PLA a pathologist at MUSC’s biorepository core (27). facility.

Lung cancer specific Q-PCR based gene array Western blotting RT2 profiler PCR array (PAHS-134C) from 1X106 A549 cells stably overexpressing either Qiagen was used for this purpose. Total RNA was SPHK2WT or SPHK2G212E were plated in a 100mm isolated from shScr and shSPHK2 tumor tissues petri dish overnight. The cells were harvested and using RNeasy kit (Qiagen) with DNase digestion centrifuged at 1300 rpm for 3.5 min and washed step and one µg of purified RNA was used for with 1X PBS (pH7.4) (Gibco). Cells were lysed cDNA synthesis using iScript cDNA synthesis kit with 1X CHAPS lysis buffer [10mM tris-HCl (Bio-Rad). PCR array was carried out using (pH7.5), 1mM MgCl2, 1mM EDTA, 0.1mM Applied Biosystems (ABI 7300) and analysis was benzamidine, 5mM beta-mercaptoethanol, 0.5% performed as per manufacturer’s instructions. CHAPS, 10% glycerol] including a protease inhibitor cocktail (Sigma-Aldrich) for 20 min on Quantitative PCR (qPCR) ice. Cell lysates were centrifuged at 12000g for 15 RNeasy (Qiagen) kit was used to isolate total min at 4oC, and supernatants were used for RNA, and 1µg of total RNA was used for Western Blotting. SDS-PAGE was carried out

9 using the Bio-Rad Criterion apparatus, followed Olympus FV10i microscope with 594- and 488- by semi-dry transfer onto a polyvinylidene nm channels for visualizing red and green difluoride (PVDF) membrane. The membrane was fluorescence. Images were taken at 60X blocked with 5% milk + 0.1% Tween-20 in 1X magnification. At least three random fields were PBS (pH7.4). Primary antibodies were used at selected for images. Caspase-3/7 activity was 1:1000 dilution overnight at 4oC, followed by measured in A549 cells and mouse embryonic rabbit or mouse secondary antibodies (Jackson fibroblasts by caspase-3 fluorometric assay kit Research Laboratories) conjugated with (BF-1100, R&D systems) as per manufacturer horseradish peroxidase at room temperature for 1 instructions. Moreover, A549 cells were pretreated hour. The rabbit secondary antibody was used at with Z-DEVD-FMK (caspase-3 specific inhibitor, 1:2500 for detecting hTERT and at 1: 5000 FMK-004, R&D systems) at 10 µM for two hours dilution for the rest of the antibodies. Actin was prior to ABC294640 treatment and caspase-3 used as an internal control for Western blotting. activity measurement. Also, caspase-3 activity was The membranses were washed either with 1X PBS measured using caspase-3 fluorometric kit with Tween-20 or 1X tris-buffered saline with following pretreatment with 5µM S1P. TUNEL 0.1% Tween-20 and developed using ECL plus assay: DeadEnd™ Colorimetric TUNEL System chemiluminescence detection kit (GE healthcare). (G7360, Promega) was used for measuring TUNEL positive cells in mouse tissues as per Downloaded from Detection of telomere damage by TIF assay manufacturer instructions. 50,000 cells were plated in each of 4-well chamber slides overnight. Following treatment with Serum starvation and cell cycle analysis by flow ABC294640 (80 µM for 8 hours), the cells were cytometry fixed in 4% paraformaldehyde in 1X PBS, pH7.4 A549 cells (1 x 106) were seeded in 100 mm http://www.jbc.org/ for 15 minutes at room temperature. The slides dishes and incubated at 37°C for 16 h before being were washed thrice with 1X PBS pH 7.4 and switched to serum free media for 48 h. Then, cells blocked with 1% goat serum in 0.3M glycine, 1% were either treated with DMSO or 80 µM BSA and 0.1% Tween-20 in 1X PBS, pH7.4 for 2 ABC294640 for 8 additional hours. Cells were by guest on May 31, 2018 hours. The cells were then incubated with primary collected by trypsinization and washed twice in antibodies that recognize g-H2AX (5µg/ml; 1X PBS. The cells were then fixed for 30 minutes ab2893, Abcam) and TRF-2 (5µg/ml; IMG-124A, at 4°C in cold 70% ethanol (added drop- wise to Imgenex) for 18 h. The slides were then washed the pellet while vortexing to ensure fixation of all thrice with 1X PBS, pH7.4 and incubated with cells and to minimize clumping). The pellets were secondary antibodies containing Alexa 488, Alexa spun at 850 g for 5 minutes, and rinsed twice with 594 and Cy5 fluorophores against g-H2AX and 1X PBS. The cells were then treated with TRF-2, respectively for 2 h at 20-25 oC. The slides ribonuclease, and 5 µg/mL propidium iodide (PI) were then washed three times with 1X PBS, pH7.4 was added before analysis by FACS. and DAPI containing mounting media was added and sealed with glass cover slips (colocalization Measurement of sphingolipids using mass resulted in either purple or yellow/orange images). spectrometry by lipidomics Images were captured using IF-CM, and Sphingolipids in cells (1 x 106) were measured colocalization was quantified using Pearson’s using mass spectrometry by lipidomics shared correlation coefficient was used to calculate resource facility at the Medical University of colocalization efficiency and P<0.05 was South Carolina, as we described previously (54). considered significant by ImageJ Fiji software The lipid abundance was normalized to inorganic (27). phosphate (Pi),

Measurement of caspase 3 activation and Ultra-structure analysis using TEM apoptosis Formalin fixed shScr and shTCF21 tumors treated Active caspase-3 was measured using cleaved with ABC294640 or vehicle control were used. caspase-3 antibody (2305-PC-020) from Trevigen After post-fixation in 2% (v/v) osmium tetroxide, Inc. Immunofluorescence was performed using specimens were embedded in Epon 812, and

10 sections were cut orthogonally to the cell and hTERT levels were measured by monolayer with a diamond knife. Thin sections immunohistochemistry. were visualized in a JEOL 1010 TEM (55,56). Xenograft-derived tumors in SCID mice Detection of senescence expressing stable shSPHK2 A549 cancer cells and mouse embryonic A549 cells stably expressing SCR-shRNA or fibroblasts (wild type, SphK1-/-, SphK2-/-) at shRNA against SPHK2 (TRCN00000036973) passage 6 were plated in 60mm petri dishes for 18 were used for the experiment. Two million cells hours followed by detection of senescence using were implanted (100 µl) into both flanks of SCID senescence-associated ß-galactosidase assay kit as mice (n=4 female SCID mice per group/ one described by the manufacturer (Cell Signaling tumor on each side of the flank) and when the Technology) (27). tumors were palpable the mice were either treated with vehicle control of with ABC294640 at 100 Measurement of senescence, aging and tumor mg/kg body weight for 21 days. Tumor volume suppression in mice was measured every alternative day using digital Mouse testes tissues at different mating calipers and treatments were performed using oral generations and tumor xenografts were fresh gavage method every day. At the end of the frozen and 5µm tissue sections were cut and experiment the mice were euthanized and tumor Downloaded from stained for senescence using senescence associated tissues were collected (27, 54-56). ß-galactosidase assay kit as described by manufacturer’s instructions (Cell Signaling Xenograft-derived tumors in SCID mice Technology). SA-ß- galactosidase activity was expressing stable shTCF21 also measured for paw tissues of mice from A549 cells stably expressing control shRNA or http://www.jbc.org/ indicated genotypes using the same kit. All animal shRNA against TCF21 (TRCN0000015564, Open experiment protocols were approved by the biosystems) were used for the experiment. Two Institutional Animal Care and Use Committee at million cells (100 µl) were implanted into both the Medical University of South Carolina (27, 54- flanks of female SCID mice (n=4 mice per group/ by guest on May 31, 2018 56). one tumor on each side of the flank) and when the tumors were palpable the mice were either treated Orthotopic lung tumor development in SCID with vehicle control of with ABC294640 at 100 mice mg/kg body weight for 21 days. Tumor volume SCID (severe combined immunodeficient) mice was measured every alternative day using digital were purchased from Harlan laboratories. Age and calipers and treatments were performed using oral sex matched mice (female SCID mice, n=7) were gavage method every day. At the end of the used for this experiment. Mice were injected (100 experiment the mice were euthanized and tumor µl) with 1 million A549-Luciferase cells through tissues were collected (27, 54-57). tail vein and were allowed to colonize lungs. Ten days after tail vein injection, the mice were Expression and purification of caspase 3. Wild- anesthetized with isoflurane and D-Luciferin was type human caspase-3 was expressed in E. coli and administered through intra peritoneal injection and purified using Ni-NTA agarose affinity resin, as luciferase bioluminescence was measured using desccribed (58,59). Briefly, C-terminal His6- Xenogen IVIS 200 Bioluminescence and tagged plasmids pET23b-Casp3-His was Fluorescence Imaging System. Once luciferase purchased from Addgene (plasmid #’s 11821 and expression was measured in the lungs of the 90090, respectively) and transformed into E. coli animals they were segregated into two groups, one strain C41(DE3). A single colony was used to group received vehicle control (50%DMSO and inoculate 50 mL of LB containing 50 µg/mL 50% PEG-200) and the other received ampicillin and grown overnight at 37 °C. This ABC294640 (100mg/Kg body weight) for 16 culture was diluted 1:2000 into 4x250 mL of LB days. At the end of the experiment the mice were media and incubated at 37 °C until the OD600 euthanized, lung tissues were collected and Ki-67 reached 3-5. Rich media was exchanged to a minimal formulation suitable for high-density

11 IPTG induction by spinning the cells gently at a flow rate of 20 µL/min in 1X PBS buffer (pH 5,000 x g for 10 min at 20-25 oC, and then 7.4) and 0.1% v/v Tween 20. Free activated resuspending in an equal volume of minimal carboxyl groups were deactivated with the media. Cultures were incubated at 37 °C until the addition of 100 µL blocking buffer (Nicoya). The OD600 increased by at least 1 unit, the incubator immobilized ligand was washed with analyte was lowered to 18 °C, and protein expression was running buffer (1X PBS, pH 7.4; 0.1% v/v Tween induced with 1 mM IPTG for 16 h at 170 rpm. 20; 0.1% w/v BSA) until a stable baseline was The cells were pelleted at 10,000 x g for 20 min achieved. Buffer-matched recombinant caspase 3 and were resuspended in lysis/binding buffer A (80 µL; 50-1000 nM) analytes were injected into (20mM HEPES pH 7.5, 1M NaCl, 20 mM the flow cell at a rate of 20 µL/min. Following a imidazole and 10 mM β-mercaptoethanol). 4-minute interaction time, the dissociation was Following sonication and centrifugation at 18,000 recorded for an additional 16 minutes. The ligand- x g for 30 min, the cleared lysate was applied to an immobilized sensor chip was regenerated between equilibrated 1 ml HisTrap FF column (GE each analyte injection using 20 mM HCl at a 150 Healthcare, catalog No. 17-5255-01), washed with µL/min flow rate. Kinetic binding analysis was 10 column volumes of wash buffer (10% Buffer B; performed with the TraceDrawer software package 20mM HEPES pH 7.5, 1M NaCl, 250 mM (Ridgeview Instruments, Uppsala, Sweden). imidazole and 10 mM β-mercaptoethanol) and Sensorgram traces were fit to a 1:1 Langmuir Downloaded from eluted by a linear gradient of imidazole from 10- model to derive association (Ka), dissociation 70% over 40 column volumes using an ÄKTA (Kd), and affinity (KD) constants. start FPLC (GE Healthcare Life Sciences). The eluted fractions were analyzed by SDS–PAGE and Statistical Analyses fractions containing caspase-3 were supplemented All data are presented as mean ± SD, and group http://www.jbc.org/ with 5 mM TCEP, concentrated to 2 mL and comparisons were performed with a two-tailed applied to a HiPrep™ 16/60 Sephacryl™ S-200 Student’s t test. Wilcoxon rank sum test and paired HR size exclusion column in SEC Buffer C (50 t-test were used to analyze small-cell and non- mM HEPES pH 7.5, 300 mM NaCl, 10 mM β- small cell tumor microarray data. Tumor xenograft by guest on May 31, 2018 mercaptoethanol 1 mM EDTA, and 10% glycerol). experiments were analyzed using two-way Fractions containing pure caspase-3 as determined ANOVA with Tukey’s posthoc test and testes by 4-20% SDS-PAGE were pooled, quantified, weights from G4 and G5 mice were analyzed and stored at -80 °C as 1 mL aliquots. using one-way ANOVA with Tukey’s posthoc test. The results were analysed using GraphPad Molecular modeling and OpenSPR for Prism 7 software and P<0.05 was considered measurements of protein-protein interactions statistically significant (27, 54-56). For the The exisiting PDB files containing NMR structure comparison of several groups, a variance analysis of p16 (2A5E) and X-ray crystallography structure (ANOVA) was carried out under normal of caspase 3 (5IBP) from RSCB (www.rscb.org) distribution assumption. were transferred into ZDOCK Server (http://zdock.umassmed.edu/). The top model was Acknowledgements then used to predict the sites of association We thank Dr. Richard Flavell (Yale University) between p16 and caspase 3, as we described (57). for providing us with the caspase3/7 knockout The protein-protein association between MEFs. We thank Dr. Sarah Spiegel (Virginia recombinant purified human p16 (Novus Commonwealth University) for providing SphK2 Biologicals, NBP2-35199) and caspase 3 was plasmids. The authors also thank Dr. Elizabeth measured using an OpenSPR localized surface Garrett-Mayer (MUSC) for her assistance with plasmon resonance (LSPR) biosensor (Nicoya Life statistical analyses and discussions in this study. Science, Inc., Kitchener, Canada), as described This work is supported by research funding from previously (60). In short, 100 µL (50 µg/ml) of the National Institutes of Health (R01-CA088932, p16Ink4a, purchased from Novus Biologicals R01-CA173687, R01-DE016572, and P01- (#NBP2-35199), was immobilized on a COOH CA203628 to BO). The core facilities utilized sensor chip (Nicoya # SEN-AU-100-12-COOH) at were constructed using support from NIH (C06

12 RR015455), Hollings Cancer Center Support TRF2-dependent T-loop formation. Cell Grant (P30 CA138313), or Center of Biomedical 155, 345-356 Research Excellence (Cobre) in Lipidomics and 6. Timashev, L. A., Babcock, H., Zhuang, Pathobiology (P30 GM103339). X., and de Lange, T. (2017) The DDR at telomeres lacking intact shelterin does not Conflict of Interest Statement require substantial chromatin The authors declare no conflict of interest. decompaction. Genes Dev 31, 578-589 7. de Lange, T. (2005) Shelterin: the protein Author Contributions complex that shapes and safeguards S.P.S. designed and performed experiments, human telomeres. Genes Dev 19, 2100- analyzed data, prepared the figures, and helped 2110 write the manuscript; M.C. designed animal 8. Rice, C., Shastrula, P. K., Kossenkov, A. breedings and prepared tissue sections from mice; V., Hills, R., Baird, D. M., Showe, L. C., K.H. prepared tissues from mice and provided Doukov, T., Janicki, S., and Skordalakes, pathological analyses of the tissue slides; C.D.S. E. (2017) Structural and functional synthesized and provided ABC294640; B.O. analysis of the human POT1-TPP1 designed experiments, analyzed data, and wrote telomeric complex. Nat Commun 8, 14928 the manuscript. 9. Counter, C. M., Hahn, W. C., Wei, W., Downloaded from Caddle, S. D., Beijersbergen, R. L., For more information Lansdorp, P. M., Sedivy, J. M., and Lipidomics: Weinberg, R. A. (1998) Dissociation http://www.hollingscancercenter.org/research/shar among in vitro telomerase activity, ed-resources/lipidomics/index.html telomere maintenance, and cellular http://www.jbc.org/ immortalization. Proc Natl Acad Sci U S A References 95, 14723-14728 1. Cohen, S. B., Graham, M. E., Lovrecz, G. 10. Gunes, C., and Rudolph, K. L. (2013) The O., Bache, N., Robinson, P. J., and role of telomeres in stem cells and cancer. by guest on May 31, 2018 Reddel, R. R. (2007) Protein composition Cell 152, 390-393 of catalytically active human telomerase 11. Valentijn, L. J., Koster, J., Zwijnenburg, from immortal cells. Science 315, 1850- D. A., Hasselt, N. E., van Sluis, P., 1853 Volckmann, R., van Noesel, M. M., 2. Gillis, A. J., Schuller, A. P., and George, R. E., Tytgat, G. A., Molenaar, J. Skordalakes, E. (2008) Structure of the J., and Versteeg, R. (2015) TERT Tribolium castaneum telomerase catalytic rearrangements are frequent in subunit TERT. Nature 455, 633-637 neuroblastoma and identify aggressive 3. Greider, C. W., and Blackburn, E. H. tumors. Nat Genet 47, 1411-1414 (1987) The telomere terminal transferase 12. Hahn, W. C., Stewart, S. A., Brooks, M. of Tetrahymena is a ribonucleoprotein W., York, S. G., Eaton, E., Kurachi, A., enzyme with two kinds of primer Beijersbergen, R. L., Knoll, J. H., specificity. Cell 51, 887-898 Meyerson, M., and Weinberg, R. A. 4. Jiang, J., Chan, H., Cash, D. D., Miracco, (1999) Inhibition of telomerase limits the E. J., Ogorzalek Loo, R. R., Upton, H. E., growth of human cancer cells. Nat Med 5, Cascio, D., O'Brien Johnson, R., Collins, 1164-1170 K., Loo, J. A., Zhou, Z. H., and Feigon, J. 13. Bodnar, A. G., Ouellette, M., Frolkis, M., (2015) Structure of Tetrahymena Holt, S. E., Chiu, C. P., Morin, G. B., telomerase reveals previously unknown Harley, C. B., Shay, J. W., Lichtsteiner, subunits, functions, and interactions. S., and Wright, W. E. (1998) Extension of Science 350, aab4070 life-span by introduction of telomerase 5. Doksani, Y., Wu, J. Y., de Lange, T., and into normal human cells. Science 279, Zhuang, X. (2013) Super-resolution 349-352 fluorescence imaging of telomeres reveals

13 14. Walsh, K. M., Codd, V., Smirnov, I. V., dynamics and senescence. Nat Struct Mol Rice, T., Decker, P. A., Hansen, H. M., Biol 20, 1199-1205 Kollmeyer, T., Kosel, M. L., Molinaro, A. 22. Alvarez, S. E., Harikumar, K. B., Hait, N. M., McCoy, L. S., Bracci, P. M., Cabriga, C., Allegood, J., Strub, G. M., Kim, E. Y., B. S., Pekmezci, M., Zheng, S., Wiemels, Maceyka, M., Jiang, H., Luo, C., Kordula, J. L., Pico, A. R., Tihan, T., Berger, M. S., T., Milstien, S., and Spiegel, S. (2010) Chang, S. M., Prados, M. D., Lachance, Sphingosine-1-phosphate is a missing D. H., O'Neill, B. P., Sicotte, H., Eckel- cofactor for the E3 ubiquitin ligase Passow, J. E., Group, E. C. T., van der TRAF2. Nature 465, 1084-1088 Harst, P., Wiencke, J. K., Samani, N. J., 23. Hannun, Y. A., and Obeid, L. M. (2008) Jenkins, R. B., and Wrensch, M. R. (2014) Principles of bioactive lipid signalling: Variants near TERT and TERC lessons from sphingolipids. Nat Rev Mol influencing telomere length are associated Cell Biol 9, 139-150 with high-grade glioma risk. Nat Genet 24. Ogretmen, B. (2018) Sphingolipid 46, 731-735 metabolism in cancer signalling and 15. Horn, S., Figl, A., Rachakonda, P. S., therapy. Nat Rev Cancer 18, 33-50 Fischer, C., Sucker, A., Gast, A., Kadel, 25. Pyne, N. J., and Pyne, S. (2010) S., Moll, I., Nagore, E., Hemminki, K., Sphingosine 1-phosphate and cancer. Nat Downloaded from Schadendorf, D., and Kumar, R. (2013) Rev Cancer 10, 489-503 TERT promoter mutations in familial and 26. Hait, N. C., Allegood, J., Maceyka, M., sporadic melanoma. Science 339, 959-961 Strub, G. M., Harikumar, K. B., Singh, S. 16. Li, Y., Zhou, Q. L., Sun, W., K., Luo, C., Marmorstein, R., Kordula, T., Chandrasekharan, P., Cheng, H. S., Ying, Milstien, S., and Spiegel, S. (2009) http://www.jbc.org/ Z., Lakshmanan, M., Raju, A., Tenen, D. Regulation of histone acetylation in the G., Cheng, S. Y., Chuang, K. H., Li, J., nucleus by sphingosine-1-phosphate. Prabhakar, S., Li, M., and Tergaonkar, V. Science 325, 1254-1257 (2015) Non-canonical NF-kappaB 27. Panneer Selvam, S., De Palma, R. M., by guest on May 31, 2018 signalling and ETS1/2 cooperatively drive Oaks, J. J., Oleinik, N., Peterson, Y. K., C250T mutant TERT promoter activation. Stahelin, R. V., Skordalakes, E., Nat Cell Biol 17, 1327-1338 Ponnusamy, S., Garrett-Mayer, E., Smith, 17. Huang, F. W., Hodis, E., Xu, M. J., C. D., and Ogretmen, B. (2015) Binding Kryukov, G. V., Chin, L., and Garraway, of the sphingolipid S1P to hTERT L. A. (2013) Highly recurrent TERT stabilizes telomerase at the nuclear promoter mutations in human melanoma. periphery by allosterically mimicking Science 339, 957-959 protein phosphorylation. Sci Signal 8, ra58 18. Vega, L. R., Mateyak, M. K., and Zakian, 28. Lewis, C.S., Voelkel-Johnson, C., and V. A. (2003) Getting to the end: Smith, C.D (2016). Suppression of c-Myc telomerase access in yeast and humans. and RRM2 expression in pancreatic Nat Rev Mol Cell Biol 4, 948-959 cancer cells by the sphingosine kinase-2 19. Montero, J. J., Lopez de Silanes, I., Grana, inhibitor ABC294640. Oncotarget 7, O., and Blasco, M. A. (2016) Telomeric 60181-60192 RNAs are essential to maintain telomeres. 29. French, K. J., Zhuang, Y., Maines, L. W., Nat Commun 7, 12534 Gao, P., Wang, W., Beljanski, V., Upson, 20. Shay, J. W. (2016) Role of Telomeres and J. J., Green, C. L., Keller, S. N., and Telomerase in Aging and Cancer. Cancer Smith, C. D. (2010) Pharmacology and Discov 6, 584-593 antitumor activity of ABC294640, a 21. Balk, B., Maicher, A., Dees, M., selective inhibitor of sphingosine kinase- Klermund, J., Luke-Glaser, S., Bender, K., 2. J Pharmacol Exp Ther 333, 129-139 and Luke, B. (2013) Telomeric RNA- 30. Venkata, J. K., An, N., Stuart, R., Costa, DNA hybrids affect telomere-length L. J., Cai, H., Coker, W., Song, J. H., Gibbs, K., Matson, T., Garrett-Mayer, E.,

14 Wan, Z., Ogretmen, B., Smith, C., and immunoblotting, immunoprecipitation and Kang, Y. (2014) Inhibition of sphingosine immunofluorescence. F1000Res 5, 2825 kinase 2 downregulates the expression of 39. Masterson, J. C., and O'Dea, S. (2007) 5- c-Myc and Mcl-1 and induces apoptosis in Bromo-2-deoxyuridine activates DNA multiple myeloma. Blood 124, 1915-1925 damage signalling responses and induces a 31. Mender, I., and Shay, J. W. (2015) senescence-like phenotype in p16-null Telomere Dysfunction Induced Foci (TIF) lung cancer cells. Anticancer Drugs 18, Analysis. Bio Protoc 5 1053-1068 32. Lakhani, S. A., Masud, A., Kuida, K., Porter, G. A., Jr., Booth, C. J., Mehal, W. 40. Bose, K., Pop, C., Feeney, B., and Clark, Z., Inayat, I., and Flavell, R. A. (2006) A.C. (2003) An uncleavable procaspase-3 Caspases 3 and 7: key mediators of mutant has a lower catalytic efficiency but mitochondrial events of apoptosis. Science an active site similar to that of mature 311, 847-851 caspase-3. Biochemistry 42, 12298-12310 33. Igarashi, N., Okada, T., Hayashi, S., Fujita, T., Jahangeer, S., and Nakamura, S. 41. Strub, G. M., Paillard, M., Liang, J., (2003) Sphingosine kinase 2 is a nuclear Gomez, L., Allegood, J. C., Hait, N. C., protein and inhibits DNA synthesis. J Biol Maceyka, M., Price, M. M., Chen, Q., Downloaded from Chem 278, 46832-46839 Simpson, D. C., Kordula, T., Milstien, S., 34. Hait, N. C., Bellamy, A., Milstien, S., Lesnefsky, E. J., and Spiegel, S. (2011) Kordula, T., and Spiegel, S. (2007) Sphingosine-1-phosphate produced by Sphingosine kinase type 2 activation by

sphingosine kinase 2 in mitochondria http://www.jbc.org/ ERK-mediated phosphorylation. J Biol interacts with prohibitin 2 to regulate Chem 282, 12058-12065 complex IV assembly and respiration. 35. Ponnusamy, S., Selvam, S. P., Mehrotra, FASEB J 25, 600-612 S., Kawamori, T., Snider, A. J., Obeid, L. M., Shao, Y., Sabbadini, R., and 42. Parham, K. A., Zebol, J. R., Tooley, K. L., by guest on May 31, 2018 Ogretmen, B. (2012) Communication Sun, W. Y., Moldenhauer, L. M., between host organism and cancer cells is Cockshell, M. P., Gliddon, B. L., Moretti, transduced by systemic sphingosine kinase P. A., Tigyi, G., Pitson, S. M., and 1/sphingosine 1-phosphate signalling to Bonder, C. S. (2015) Sphingosine 1- regulate tumour metastasis. EMBO Mol phosphate is a ligand for peroxisome Med 4, 761-775 proliferator-activated receptor-gamma that 36. Smith, L. T., Lin, M., Brena, R. M., Lang, regulates neoangiogenesis. FASEB J 29, J. C., Schuller, D. E., Otterson, G. A., 3638-3653 Morrison, C. D., Smiraglia, D. J., and Plass, C. (2006) Epigenetic regulation of 43. Richards, K. L., Zhang, B., Sun, M., the tumor suppressor gene TCF21 on Dong, W., Churchill, J., Bachinski, L. L., 6q23-q24 in lung and head and neck Wilson, C. D., Baggerly, K. A., Yin, G., cancer. Proc Natl Acad Sci U S A 103, Hayes, D. N., Wistuba, II, and Krahe, R. 982-987 (2011) Methylation of the candidate 37. Pont, A. R., Sadri, N., Hsiao, S. J., Smith, biomarker TCF21 is very frequent across a S., and Schneider, R. J. (2012) mRNA spectrum of early-stage nonsmall cell lung decay factor AUF1 maintains normal cancers. Cancer 117, 606-617 aging, telomere maintenance, and 44. Tessema, M., Willink, R., Do, K., Yu, Y. suppression of senescence by activation of Y., Yu, W., Machida, E. O., Brock, M., telomerase transcription. Mol Cell 47, 5- Van Neste, L., Stidley, C. A., Baylin, S. 15 B., and Belinsky, S. A. (2008) Promoter 38. Neubauer, H. A., and Pitson, S. M. (2016) methylation of genes in and around the Validation of commercially available sphingosine kinase 2 antibodies for use in

15 candidate lung cancer susceptibility locus Saltness, R. A., Jeganathan, K. B., 6q23-25. Cancer Res 68, 1707-1714 Verzosa, G. C., Pezeshki, A., Khazaie, K., 45. Arab, K., Park, Y. J., Lindroth, A. M., Miller, J. D., and van Deursen, J. M. Schafer, A., Oakes, C., Weichenhan, D., (2016) Naturally occurring p16(Ink4a)- Lukanova, A., Lundin, E., Risch, A., positive cells shorten healthy lifespan. Meister, M., Dienemann, H., Dyckhoff, Nature 530, 184-189 G., Herold-Mende, C., Grummt, I., Niehrs, 52. Baker, D. J., Wijshake, T., Tchkonia, T., C., and Plass, C. (2014) Long noncoding LeBrasseur, N. K., Childs, B. G., van de RNA TARID directs demethylation and Sluis, B., Kirkland, J. L., and van activation of the tumor suppressor TCF21 Deursen, J. M. (2011) Clearance of via GADD45A. Mol Cell 55, 604-614 p16Ink4a-positive senescent cells delays ageing-associated disorders. Nature 479, 46. Venant, H., Rahmaniyan, M., Jones, E.E., 232-236 Lu, P., Lilly, M.B., Garrett-Mayer, E., 53. Chang, J., Wang, Y., Shao, L., Laberge, R. Drake, R.R., Kraveka, J.M., Smith, C.D., M., Demaria, M., Campisi, J., and Voelkel-Johnson, C. (2015) The Janakiraman, K., Sharpless, N. E., Ding, sphingosine kinase 2 inhibitor S., Feng, W., Luo, Y., Wang, X., Aykin- ABC294640 reduces the growth of Burns, N., Krager, K., Ponnappan, U., Downloaded from prostate cancer cells and results in Hauer-Jensen, M., Meng, A., and Zhou, D. accumulation of dihydroceramides in vitro (2016) Clearance of senescent cells by and in vivo. Mol Cancer Ther. 14, 2744- ABT263 rejuvenates aged hematopoietic 2752 stem cells in mice. Nat Med 22, 78-83 54. Gencer, S., Oleinik, N., Kim, J., Panneer http://www.jbc.org/ 47. Britten, C.D., Garrett-Mayer, E., Chin, Selvam, S., De Palma, R., Dany, M., S.H., Shirai, K., Ogretmen, B., Bentz, Nganga, R., Thomas, R.J., Senkal, C.E., T.A., Brisendine, A., Anderton, K., Howe, P.H., and Ogretmen, B. (2017) TGF-β receptor I/II trafficking and

Cusack, S.L., Maines, L.W., Zhuang, Y., by guest on May 31, 2018 Smith, C.D., and Thomas, M.B. (2017) A signaling at primary cilia are inhibited by Phase I study of ABC294640, a first-in- ceramide to attenuate cell migration and class sphingosine kinase-2 inhibitor, in tumor metastasis. Sci Signal. 10. pii: patients with advanced solid tumors. Clin eaam7464 Cancer Res. 23, 4642-4650 55. Dany, M., Gencer, S., Nganga, R., 48. Blasco, M. A., Lee, H. W., Hande, M. P., Thomas, R. J., Oleinik, N., Baron, K. D., Samper, E., Lansdorp, P. M., DePinho, R. Szulc, Z. M., Ruvolo, P., Kornblau, S., A., and Greider, C. W. (1997) Telomere Andreeff, M., and Ogretmen, B. (2016) shortening and tumor formation by mouse Targeting FLT3-ITD signaling mediates cells lacking telomerase RNA. Cell 91, ceramide-dependent mitophagy and 25-34 attenuates drug resistance in AML. Blood 49. Wang, Y., Sharpless, N., and Chang, S. 128, 1944-1958 (2013) p16(INK4a) protects against 56. Sentelle, R. D., Senkal, C. E., Jiang, W., dysfunctional telomere-induced ATR- Ponnusamy, S., Gencer, S., Selvam, S. P., dependent DNA damage responses. J Clin Ramshesh, V. K., Peterson, Y. K., Invest 123, 4489-4501 Lemasters, J. J., Szulc, Z. M., Bielawski, 50. Jacobs, J. J., and de Lange, T. (2005) J., and Ogretmen, B. (2012) Ceramide p16INK4a as a second effector of the targets autophagosomes to mitochondria telomere damage pathway. Cell Cycle 4, and induces lethal mitophagy. Nat Chem 1364-1368 Biol 8, 831-838 51. Baker, D. J., Childs, B. G., Durik, M., 57. Thomas, R. J., Oleinik, N., Panneer Wijers, M. E., Sieben, C. J., Zhong, J., Selvam, S., Vaena, S. G., Dany, M., Nganga, R. N., Depalma, R., Baron, K. D.,

16 Kim, J., Szulc, Z. M., and Ogretmen, B. independent experiments, analysed by student’s t- (2017) HPV/E7 induces chemotherapy- test,n=3, *P=0.025, **P=0.042, ***P=0.66), mediated tumor suppression by ceramide- respectively, were measured in A549 xenograft- dependent mitophagy. EMBO Mol Med 9, derived tumors. Scale bars represent 100 µm. (E- 1030-1051 F) Telomere damage (E) in A549 xenograft- 58. Hwang, D., Kim, S.A., Yang, E.G., Song, derived tumors expressing Scr- or SPHK2- H.K., and Chung, H.S. A facile method to shRNAs in the absence/presence of ABC294640 prepare large quantities of active caspase- were measured using TIF assay. Quantification of 3 overexpressed by auto-induction in the images were performed using Image J (right C41(DE3) strain. (2016) Protein Expr panel). Data are means ± SD from three Purif 126, 104–108 independent experiments, analysed by student’s t- 59. Sivashanmugam, A., Murray, V., Cui, C., test (n=3, *P=0.006, **P=0.013, ***P=0.003). (F) Zhang, Y., Wang, J., and Li, Q. Practical Senescence in A549 derived tumors expressing protocols for production of very high yields Scr-or SPHK2-shRNAs in the absence/presence of of recombinant proteins using Escherichia ABC294640 were measured using SA-beta-gal coli. (2009) Protein Sci 18, 936–948 staining. Scale bars represent 100 µm. 60. McGurn, L.D., Moazami-Goudarzi, M., White, S.A., Suwal, T., Brar, B., Tang, Fig. 2. SphK2 targeting using ABC294640 Downloaded from J.Q., Espie, G.S., and Kimber, M.S. (2016) induces caspase 3-dependent apoptosis. (A) The structure, kinetics and interactions of Effects of ABC294640 on apoptosis induction the β-carboxysomal β-carbonic anhydrase, were measured in immortalized MEFs isolated CcaA. Biochem J 473, 4559-4572 from Casp3+/-/Casp7+/-, Casp3-/-/Casp7+/-, Casp3+/- /Casp7-/-, and Casp3-/-/Casp7-/- mice were http://www.jbc.org/ Figure Legends measured compared to vehicle-treated controls using caspase 3/7 activity assay compared to Fig. 1. Inhibition of SphK2 attenuates tumor vehicle-treated controls. Data are means ± SD growth in vivo. (A) Effects of shRNA-dependent from three independent experiments, analysed by by guest on May 31, 2018 silencing of SPHK2 on A459 xenograft-derived student’s t-test (n=3, *P=0.006, **P=0.013, tumor growth in the presence/absence of ***P=0.003, ****P=0.003). (B) Telomere ABC294640 (n=4 mice per group, each containing damage in response to ABC294640-mediated two tumors on both flanks) were measured in SphK2 inhibition was measured in MEFs isolated SCID mice for 21 days (upper panel). Tumor from WT and Casp3-/-/Casp7+/- mice using TIF growth was measured every alternative day for assay. Quantification of images were performed upto 21 days by calipers. After the growth study using Image J (right panel). Data are means ± SD was completed, the tumors were removed and from three independent experiments, analysed by measured ex vivo (lower panel). Two way student’s t-test (n=3, *P=0.03, **P=0.042). Scale ANOVA with Tukey’s posthoc test was bars represent 100 µm. (C) Effects of caspase 3 performed. shScr Vs ShScr+ABC(*P<0.001, inhibition using Z-DEVD on apoptosis in response F=22.76, DF=27), shScr Vs shSK2(**P<0.001, to SPHK2 inhibition by ABC294640 were F=35.48, DF=9), shSK2 versus shSK2 + ABC measured using caspase 3/7 activity assay in A549 (***P=0.997). (B) ShRNA-mediated knockdown cells compared to vehicle-treated controls. Data of ISPHK2 mRNA in A549 xenograft-derived are means ± SD from three independent tumors was confirmed using qPCR after the tumor experiments, analysed by student’s t-test (n=3, growth measurements were completed. Data are *P=0.042, **P=0.033). (D) Ectopic expression of means ± SD from three independent experiments, WT-SphK2-V5 and SphK2G212E-V5 compared to analysed by student’s t-test (n=3, *P=0.039, empty vector-V5 in A549 cells transfected with **P=0.014). (C-D) Effects of shRNA-mediated Scr- or SphK2-shRNAs were confirmed by knockdown and/or ABC294640-dependent Western blotting using anti-V5 antibody (upper inhibition of SPHK2 on hTERT expression (C) or panel). Actin was measured as a loading control apoptosis (D) measured using IHC or TUNEL (lower panel). (E) Effects of ectopic expression of staining (Data are means ± SD from three WT-SphK2-V5 and SphK2G212E-V5 on caspase 3

17 activation were measured by immunofluorescence student’s t-test (n=3, *P=0.031, **P=0.605). Scale using anti-caspase 3 antibody that detects activated bars represent 100 µm. caspase 3 compared to empty vector-V5 in A549 cells. Images were quantified using ImageJ (right Fig. 4. S1P-hTERT binding protects caspase 3 panel). Data are means ± SD from three activation in response to ABC294640 by independent experiments, analysed by student’s t- phospho-mimicking of hTERT by S1P at S921. test (n=3, *P=0.003, **P=0.472, ***P=0.0002). (A-B) Effects of ectopic expression of WT- Scale bars represent 100 µm. (F) Effects of hTERT, hTERTD684A, hTERTS921D and exogenous S1P exposure (using BSA-conjugated hTERTS921D-D684A on caspase 3 activation in the S1P, 5µM) on caspase 3 activation in the absence/presence of ABC294640 were measured absence/presence of ABC294640 were measured in GM847 cells using immunofluorescence (A). in A549 cells compared to vehicle-treated Images were quantified by ImageJ (B). Data are controls. Data are means ± SD from three means ± SD from three independent experiments, independent experiments, analysed by student’s t- analysed by student’s t-test (n=3, *P=0.052, test (n=3, *P=0.008, **P=0.006). **P=0.010). Scale bars represent 100 µm. (C) Association of S1P with S921D-hTERT and Fig. 3. Silencing SPHK2 or hTERT induces S921D-D684A-hTERT compared to vector- telomere damage in A549 cells. (A) Effects of transfected GM847 cells was measured by PLA Downloaded from genetic loss of SphK2 in inducing telomere using anti-S1P (Sphingomab) and anti-hTERT damage were measured using TIF assay in MEFs antibodies. Data are means ± SD from three isolated from SphK2-/- compared to WT mice by independent experiments, analysed by student’s t- immunofluorescence. Images were quantified test (n=3, *P=0.00064, **P=0.0028, using ImageJ (right panel). Data are means ± SD ***P=0.0028). Scale bars represent 100 µm. (D) http://www.jbc.org/ from three independent experiments, analysed by Ectopic expression of WT-, D684A-, S921A, student’s t-test (n=3, *P=0.001). Scale bars S921A/D684A-, S921D-, and S921D-D684A- represent 100 µm. (B-C) Effects of shRNA- hTERT proteins in the presence/absence of mediated knockdown of SPHK2 (B) or hTERT in ABC294640 in GM847 cells were confirmed by by guest on May 31, 2018 A549 cells on telomere damage in the Western blotting using anti-hTERT antibody. absence/presence of ABC294640 (C) were Vector-transfected cell extracts were used as measured using TIF assay by negative controls (right panel). Actin was used as immunofluorescence. Images were quantified a loading control (lower panel). Images represent using ImageJ (upper right panel). Data are means at least three independent studies. ± SD from three independent experiments, analysed by student’s t-test(n=3,*P=0.03). Sh- Fig. 5. ABC294640-mediated telomere damage RNA-mediated knockdown of hTERT mRNA was is prevented by the S1P-hTERT complex due to confirmed using qPCR compared to Scr-shRNA- phospho-mimicking function of S1P at S921 of transfected controls (lower left panel), data are hTERT. (A-B) Effects of ectopic expression of means ± SD from three independent experiments, WT-, D684A-, S921A, S921A-D684A-, S921D-, analysed by student’s t-test(n=3,*P=0.047) and and S921D-D684A-hTERT proteins on telomere TIF colocalization was quantified. Data are means damage in the presence/absence of ABC294640 in ± SD from three independent experiments, GM847 cells were detected by analysed by student’s t-test (n=3, *P=0.030, immunofluorescence using TIF assay (A). Scale **P=0.003, ***P=0.108). Scale bars represent 100 bars represent 100 µm. Images were quantified by µm. (D) Effects of ABC294640 or cisplatin on ImageJ (B). Data are means ± SD from three p53-dependent DNA damage was measured by independent experiments, analysed by student’s t- immunofluorescence using anti-gH2AX and anti- test (n=3, *P=0.046, **P=0.037, ***P=0.033). p53 antibodies in A549 cells compared to vehicle (C) Effects of ectopically expressed WT-, D684A- (DMSO)-treated controls. Images were quantified , S921A, S921A-D684A-, S921D-, and S921D- using ImageJ (right panel). Data are means ± SD D684A-hTERT proteins on GM847 cell viability from three independent experiments, analysed by in the absence/presence of ABC294640 were measured by trypan blue exclusion assay. Data are

18 means ± SD from three independent experiments, SphK2-/- (**P<0.0001) with F=18.91 and DF=76. analysed by student’s t-test (n=3, *P=0.046, (C) Spleen, skin and testes tissues from WT and **P=0.051, ***P=0.044). SphK2-/- mice at generation 5 were isolated and examined after H&E staining. (D) Apoptosis was Fig. 6. Effects of TCF21 silencing on detected by TUNEL staining in testes tissues ABC294640-mdiated tumor suppression in isolated from WT and SphK2-/- mice at generation mice. (A-D) Effects of shRNA-dependent 4 (G4, upper panel). Protein abundance of mTERT silencing of TCF21 on A459 xenograft-derived and tcf21 were detected by IHC using antitumor growth in the presence/absence of mTERT and anti-tcf21 antibodies in testes tissues ABC294640 (n=4 mice per group, each containing isolated from WT, SphK1-/- and SphK2-/- mice at two tumors on both flanks) were measured in generation 4 (G4, lower panels). (E) Telomere SCID mice for 21 days. Tumor growth was damage in testes isolated from G4 of WT, SphK1-/- measured every alternative day for up to 21 days and SphK2-/- mice was measured using TIF assay. by calipers. After the growth study was completed, Scale bars represent 100 µm. All mice were the tumors were removed and measured ex vivo. between 6-8 weeks old. Two-way ANOVA with Tukey’s posthoc test was performed. shScr Vs ShScr+ABC(*P<0.001, Fig. 8. Genetic loss of SphK2 results in F=6.195, DF=30), shScr Vs shTCF21(**P=0.275), accelerated p16-dependent senescence but not Downloaded from shTCF21 Vs shTCF21+ABC(***P=0.872) (A). apoptosis via increased telomere damage in Protein abundance of hTERT (n=3, *P=0.0478, non-cancerous fibroblasts. (A) Senescence in **P=0.0346 by student’s t-test) (B), activation of testes tissues obtained from WT and SphK2-/- mice caspase 3 (C), and induction of apoptosis (D) in at generations 4, 5 and 6 was measured by SA- tumors were measured by Western blotting using beta-gal staining. Scale bars represent 100 µm. (B) http://www.jbc.org/ anti-hTERT antibody, IHC using anti-active Abundance of p16 mRNA was measured by qPCR caspase 3 antibody, or TEM, respectively. (E) in MEFs isolated from WT, SphK1-/- and SphK2-/- ShRNA-mediated knockdown of TCF21 using two mice at generations 6 in the absence/presence of distinct shRNAs (#1 or #2) in the Scr- or p16-shRNAs. Data are means ± SD from by guest on May 31, 2018 absence/presence of ABC294640 was measured three independent experiments, analysed by by qPCR in A549 cells. Data are means ± SD from student’s t-test (n=3, *P=0.0402, **P=0.0019) (C- three independent experiments, analysed by D) Effects of p16 knockdown on telomere damage student’s t-test (n=3, *P=0.015, **P=0.422, (C), Data are means ± SD from three independent **P=0.054). (F-G) Effects of TCF21 silencing experiments, analysed by student’s t-test (n=3, using shRNA on telomere damage in the absence *P=0.006, **P=0.013) or caspase 3 activation (D) (F) or presence (G) of ABC294640 (ABC) were were measured using immunofluorescence in measured by TIF assay using anti-gamma-H2AX MEFs obtained from WT, SphK1-/-, and SphK2-/- and TRF-2 antibodies by immunofluorescence. mice at generation 6. Scr-shRNA transfected Data represent at least three independent SphK2-/- MEFs were used as controls. Images experiments. Scr-shRNA and/or vehicle-treated were quantified using ImageJ (right panels). Data cells were used as controls. Images were are means ± SD from three independent quantified using Image J. Scale bars represent 100 experiments, analysed by student’s t-test (n=3, µm. *P=0.001).

Fig. 7. Genetic loss of SphK2 results in Fig. 9. Reconstitution of p16 induces senescence accelerated senescence and aging in mice. (A) and prevents apoptosis in cancer cells in Senescence was detected using SA-beta-gal response to ABC294640-mediated SPHK2 staining in the paws of WT, SphK1-/- and SphK2-/- inhibition and telomere damage. (A) Effects of mice at generation 5. (B) Testes were isolated and stable expression of p16 on caspase 3 activation in weighed from WT, SphK1-/- and SphK2-/- mice at A549 cells transfected with Scr- or SPHK2- generations 4 and 5 (G4-G5). One-way ANOVA shRNAs were measured by immunofluorescence. with Tukey’s posthoc analysis test was performed Images were quantified by ImageJ (lower panel). (WT Vs SphK2-/- (*P<0.0001) and SphK1-/- Vs Data are means ± SD from three independent

19 experiments, analysed by student’s t-test (n=3, using the PLA software as described by the *P=0.001, **P=0.0005, ***P=0.0004). manufacturer. Data are means ± SD from three Expression levels of p16 protein in stable independent experiments, analysed by student’s t- shSPHK2 Vs shScr A549 cells were confirmed by test (n=3, *P=0.0349). Successful knockdown of Western blotting using anti-p16 antibody (Upper p16 using shRNA was confirmed by Q-PCR in right panel). Actin was used as a loading control H1341 cells compared to Scr-shRNA-transfected (Lower right panel). (B-C) Effects of stable cells (right panel). Data are means ± SD from expression of p16 on senescence in response to three independent experiments, analysed by shRNA-knockdown (B) or ABC294640-mediated student’s t-test (n=3, *P=0.004). (B-C) Interaction inhibition (C) of SPHK2 were measured by SA- and co-localization of p16 and caspase 3 in beta-gal staining in A549 cells compared to Scr- response to SPHK2 inhibitor ABC294640 (ABC) transfected and/or vehicle-treated controls. (D-E) in the absence/presence of ectopic expression of Effects of stable expression of p16 on telomere WT-p16 (p16) and/or shRNA-dependent damage (D), Images were quantified by ImageJ, knockdown of SPHK2 (shSPHK2) were measured Data are means ± SD from three independent using PLA. Data are means ± SD from three experiments, analysed by student’s t-test (n=3, independent experiments, analysed by student’s t- *P=0.006, **P=0.002) or caspase 3 activation (E) test (n=3, *P=0.045, **P=0.094, ***P=0.005) (B) in A549 cells in the absence/presence of or confocal microscopy and immunofluorescence Downloaded from ABC294640 were measured using TIF assay or (C) in A549 cells. Data are means ± SD from three caspase 3 activation assay by immunofluorescence independent experiments, analysed by student’s t- and confocal microscopy. Images were quantified test (n=3, *P=0.484, **P=0.001, ***P=0.0009, by ImageJ (right panels). Data are means ± SD ****P=0.001). Vehicle-treated, vector-only or from three independent experiments, analysed by Scr-shRNA-transfected cells were used as http://www.jbc.org/ student’s t-test (n=3, *P=0.0006, **P=0.0005, controls. Data represent at least three independent ***P=0.0004). (F) ShRNA-mediated knockdown studies. Quantification of PLA images were of p16 was confirmed using qPCR in H1341 cells performed using the PLA software as described by compared to Scr-shRNA-transfected controls. the manufacturer. Quantification of co-localization by guest on May 31, 2018 Data are means ± SD from three independent was performed using Image J. ShRNA-mediated experiments, analysed by student’s t-test (n=3, knockdown of SPHK2, and ectopic expression of *P=0.00005). (G) Caspase 3 activation was WT-p16 were confirmed using Q-PCR(upper measured by immunofluorescence using anti- panels) and Western blotting respectively (lower caspase 3 antibody that detects active-caspase 3 in panels). Data are means ± SD from three H1341 with/without p16 knockdown compared to independent experiments, analysed by student’s t- Scr-shRNA-transfected controls in the test (n=3, *P=0.049). (D) Molecular modeling for absence/presence of ABC294640 (right panel). the predicted interaction between p16 and caspase Effects of ectopic expression of p16 on caspase 3 3 was generated using already exisiting NMR activation in the presence/absence of ABC294640 structure of p16 and crystal structure caspase 3 as was measured in A549 cells compared to vector- described in Materials and Methods. Ser152 of p16 transfected controls (left panel). (H) Expression (turquoise) and Gly251 of caspase 3 (grey) were levels of p16 protein in different lung cancer cell among key residues for the interaction between lines were confirmed by Western Blotting using these two proteins. (E) Association between p16 anti-p16 antibody (upper panel). Actin was used as and caspase 3 was measured by PLA in A549 cells a loading control (lower panel). expressing WT-p16 or p16S152A in the absence/presence of ABC294640. Vector-only- Fig. 10. Interaction between p16 and caspase 3 transfected and vehicle-treated cells were used as prevents apoptosis and induces senescence in controls. Data are means ± SD from three response to telomere damage. (A) Association independent experiments, analysed by student’s t- between endogenous p16 and caspase 3 was test (n=3, *P=0.010, **P=0.009, ***P=0.009). measured using PLA in H1341 cells in the Quantification of PLA images were performed absence/presence of Scr- or p16-shRNAs. using the PLA software as described by the Quantification of PLA images were performed manufacturer. (F-G) Effects of ABC294640

20 exposure on caspase 3 activation were measured (A) Chemical structure of ABC294640. (B-C) by confocal microscopy and immunofluorescence Effects of inhibition of SPHK2 using ABC294640 using anti-caspase 3 antibody that detects activated on A549-luciferase xenograft-derived lung tumors caspase 3 in A549 cells expressing WT-p16 or (100 mg/kg for 14 days) established orthotopically p16S152A (F). Data are means ± SD from three in lungs of SCID mice (n=7 mice per group) after independent experiments, analysed by student’s t- tail vein injections (1 X106 cells). Images were test (n=3, *P=0.0001, **P=0.643, ***P=0.0006). obtained and quantified (B-C) using Xenogen Vector-only-transfected and vehicle-treated cells IVIS 200 Bioluminescence and Fluorescence were used as controls. Quantification of co- Imaging System. *P=0.05 by student t-test. (D) localization was performed using Image J. Effects of ABC294640 on cell proliferation and Expression of WT-p16 and p16S152A was hTERT abundance in lung tumor tissues (shown in confirmed by Western blotting (G). Vector- B-C) were detected by IHC using anti-Ki-67 and transfected cell extracts were used as controls. (H) anti-hTERT antibodies. Scale bars represent 100 Association between p16 and caspase 3 was µm. measured by PLA in testes tissues obtained from WT (SphK1+/+/SphK2+/+), SphK1-deficient Supplemental Figure 2. Increased SPHK2 and (SphK1-/-) or SphK2-deficient (SphK2-/-) mice at hTERT protein abundance is detected in lung generation 6. Data are means ± SD from three tumor tissues obtained from patients. (A) Downloaded from independent experiments, analysed by student’s t- Expression of SPHK2 in lung tumor tissues test (n=3, *P=0.002, **P=0.001). Quantification obtained from patients (n=5) compared to their of PLA images were performed using the PLA matched non-cancerous adjacent lung tissues software as described by the manufacturer. Scale (n=5) were detected by IHC using anti-SPHK2 bars represent 100 µm. antibody. Scale bars represent 100 µm. (B) http://www.jbc.org/ Expression of SPHK2 (left panel) and hTERT Fig. 11. Graphical summary. Our data suggest (right panel) in NSCLC tissues (n=48) in TMAs that SPHK2-generated S1P binds and stabilizes containing lung tumors (T), adjacent non- TERT involving its D684 residue via mimicking cancerous lung tissues (AN, n=48), and normal by guest on May 31, 2018 TERT phosphorylation at S921, which protects lung tissues (n=4) was detected by IHC using anti- telomere damage, and results in delayed SPHK2 and anti-hTERT antibodies. Scale bars senescence and protection from apoptosis. Our represent 100 µm. Statistical analyses of the data data also suggest that targeting/inhibition of the were performed using paired t-test (P<0.0001). SPHK2/S1P/telomerase axis results in telomere (C) Co-expression of SPHK2 and hTERT in damage, which signals TCF21-dependent caspase NSCLC tumors (n=96) in TMAs (shown in B) was 3 activation in cancer cells or immortalized MEFs, analyzed using paired t-test. (D) Expression of or p16-dependent senescence and accelerated SPHK2 (left panel) and hTERT (right panel) in aging phenotype in non-cancerous primary lung SCLC tissues (n=40) in TMAs containing lung fibroblasts, primary MEFs, or non-cancerous tumors (T), and normal lung tissues (N, n=10) was tissues, such as testes, of subsequent generations detected by IHC using anti-SPHK2 and anti- of SphK2-/- mice. Thus, these data reveal that p16 hTERT antibodies. Statistical analyses of the data abundance induces senescence and prevents (shown in D) were performed using the Wilcoxon caspase-3-dependent apoptosis in response to ran sum test, *P=0.0012, **P=0.031 (right telomere damage via p16-caspase 3 interaction in panels). response to targeting the SPHK2/S1P/telomerase axis. Supplemental Figure 3. Inhibition or shRNA- mediated knockdown of SPHK2 induces telomere damage and caspase 3 activation in SUPPLEMENTAL FIGURE LEGENDS growth arrested A549 cells. (A-B) Effects of ABC294640 (ABC) (A) or shRNA-mediated Supplemental Figure 1. Targeting SPHK2 using knockdown of SPHK2 (B) on telomere damage ABC294640 inhibits orthotopic A549 xenograft- (upper panels) using TIF assay, and caspase 3 derived lung tumors and hTERT abundance. activation using anti-caspase 3 antibody that

21 detects activated caspase 3 (lower panels) in respectively) were confirmed by qPCR in A549 growth arrested A549 cells in response to serum cells compared to Scr-shRNA-transfected controls. starvation were measured by immunofluorescence. Data are means ± SD from three independent Data represent at least three independent experiments, analyzed by student’s t-test (n=3, experiments. (C-E) Cell cycle profiles of A549 *P=0.038). (D) Effects of ectopic expression of cells in response to serum starvation (C-D) WT-hTERT or hTERTD684A on TCF21 mRNA in compared to controls, cells grown in complete response to shRNA-mediated knockdown of media with serum (E), with/without ABC294640, SPHK2 compared to vector-and/or Scr-shRNA- ABC (C) or shRNA-SPHK2, SK2 (D) were transfected control A549 cells were measured by measured using flow cytometry. Data represent qPCR. Data are means ± SD from three three independent experiments, and error bars independent experiments, analyzed by student’s t- represent standard deviation. test. (n=3, *P=0.0481, **P=0.0174, ***P=0.0267) (E) Knockdown of TCF21 using shRNA was Supplemental Figure 4. Effects of SPHK2 confirmed in A549 cells in the presence of knockdown on ceramide, dihydro-ceramide ABC294640 compared to Scr-transfected controls and S1P were measured by lipidomics. (A-B) by qPCR. Data are means ± SD from three Effects of shRNA-mediated knockdown of independent experiments, analyzed by student’s t- SPHK2, SK2 mRNA, measured by qPCR (A) on test (n=3, *P=0.029, **P=0.035). (F) Effects of Downloaded from sphingolipids (n=4, *P=0.02), such as ceramide, shRNA-mediated knockdown of TCF21 on dihydro-C16-ceramide, sphingosine, dihydro- caspase 3 activation in the absence/presence of sphingosine, dihydro-S1P, or S1P were measured ABC294640 were measured by using mass spectrometry-based lipidomics (B) in immunofluorescence using anti-caspase 3 antibody serum starved and G0/G1 arrested A549 cells that detects activated caspase 3 in A549 cells. http://www.jbc.org/ (n=6). Data are means ± SD (std dev) analyzed by Images were quantified using ImageJ (bottom student’s t-test. panel). Data are means ± SD from three independent experiments, analyzed by student’s t- Supplemental Figure 5. ABC294640 induces test (n=3, *P=0.022, **P=0.036, ***P=0.019). by guest on May 31, 2018 telomere damage but not caspase 3 activation in Scale bars represent 100 µm. primary non-cancerous human lung fibroblasts. Effects of ABC294640 on telomere Supplemental Figure 7. Recombinant purified damage (upper panel) and caspase 3 activation p16 and caspase 3 proteins interact with each (lower panel) in human non-cancerous lung other in vitro. (A) Association between epithelial cells were measured using TIF assay and recombinant purified p16 with caspase 3 (A) or anti-caspase 3 antibody that recognizes activated mutant (pro) caspase 3 (B) was measured using caspase 3, respectively, were measured by SPR. (B) Purified caspase 3 protein is visualized immunofluorescence. Data represent at least three by SDS-PAGE and Coomassie blue staining. independent experiments.

22 Downloaded from http://www.jbc.org/ by guest on May 31, 2018 Downloaded from http://www.jbc.org/ by guest on May 31, 2018 Downloaded from http://www.jbc.org/ by guest on May 31, 2018 Downloaded from http://www.jbc.org/ by guest on May 31, 2018 Downloaded from http://www.jbc.org/ by guest on May 31, 2018 Downloaded from http://www.jbc.org/ by guest on May 31, 2018 Downloaded from http://www.jbc.org/ by guest on May 31, 2018 Downloaded from http://www.jbc.org/ by guest on May 31, 2018 Downloaded from http://www.jbc.org/ by guest on May 31, 2018 Downloaded from http://www.jbc.org/ by guest on May 31, 2018 Downloaded from http://www.jbc.org/ by guest on May 31, 2018 Balance between senescence and apoptosis is regulated by telomere damage−induced association between p16 and caspase-3 Shanmugam Panneer Selvam, Braden M Roth, Rose Nganga, Jisun Kim, Marion A Cooley, Kristi L. Helke, Charles D. Smith and Besim Ogretmen J. Biol. Chem. published online May 10, 2018

Access the most updated version of this article at doi: 10.1074/jbc.RA118.003506

Alerts: • When this article is cited • When a correction for this article is posted

Click here to choose from all of JBC's e-mail alerts Downloaded from http://www.jbc.org/ by guest on May 31, 2018 Supplemental Figures

Balance between senescence and apoptosis is regulated by telomere damage–induced association between p16 and caspase-3

Shanmugam Panneer Selvam1,2, Braden M. Roth1,2, Rose Nganga1,2, Jisun Kim1,2, Marion A. Cooley3,#, Kristi Helke4, Charles D. Smith5, and Besim Ogretmen1,2*

#Current Address: Department of Oral Biology, Augusta University, Augusta, GA 30912. *Address correspondence to: [email protected]

List of Supplemental Figures

Supplemental Figure 1. Targeting SPHK2 using ABC294640 inhibits orthotopic A549 xenograft-derived lung tumors and hTERT abundance.

Supplemental Figure 2. Increased SPHK2 and hTERT protein abundance is detected in lung tumor tissues obtained from patients.

Supplemental Figure S3. Inhibition or shRNA-mediated knockdown of SPHK2 induces telomere damage and caspase 3 activation in growth arrested A549 cells.

Supplemental Figure S4. Effects of SPHK2 knockdown on ceramide, dihydro-ceramide and S1P were measured by lipidomics.

Supplemental Figure S5. ABC294640 induces telomere damage but not caspase 3 activation in primary non-cancerous human lung fibroblasts.

Supplemental Figure 6. Targeting SPHK2 induces TCF21-mediated apoptosis and lung tumor suppression.

Supplemental Figure S7. Recombinant purified p16 and caspase 3 proteins interact with each other in vitro.

SUPPLEMENTAL FIGURE LEGENDS

Supplemental Figure 1. Targeting SPHK2 using ABC294640 inhibits orthotopic A549 xenograft- derived lung tumors and hTERT abundance. (A) Chemical structure of ABC294640. (B-C) Effects of inhibition of SPHK2 using ABC294640 on A549-luciferase xenograft-derived lung tumors (100 mg/kg for 14 days) established orthotopically in lungs of SCID mice (n=7 mice per group) after tail vein injections (1 X106 cells). Images were obtained and quantified (B-C) using Xenogen IVIS 200 Bioluminescence and Fluorescence Imaging System. *P=0.05 by student t-test. (D) Effects of ABC294640 on cell proliferation and hTERT abundance in lung tumor tissues (shown in B-C) were detected by IHC using anti-Ki-67 and anti-hTERT antibodies. Scale bars represent 100 µm.

Supplemental Figure 2. Increased SPHK2 and hTERT protein abundance is detected in lung tumor tissues obtained from patients. (A) Expression of SPHK2 in lung tumor tissues obtained from patients (n=5) compared to their matched non-cancerous adjacent lung tissues (n=5) were detected by IHC using

1 anti-SPHK2 antibody. Scale bars represent 100 µm. (B) Expression of SPHK2 (left panel) and hTERT (right panel) in NSCLC tissues (n=48) in TMAs containing lung tumors (T), adjacent non-cancerous lung tissues (AN, n=48), and normal lung tissues (n=4) was detected by IHC using anti-SPHK2 and anti-hTERT antibodies. Scale bars represent 100 µm. Statistical analyses of the data were performed using paired t-test (P<0.0001). (C) Co-expression of SPHK2 and hTERT in NSCLC tumors (n=96) in TMAs (shown in B) was analyzed using paired t-test. (D) Expression of SPHK2 (left panel) and hTERT (right panel) in SCLC tissues (n=40) in TMAs containing lung tumors (T), and normal lung tissues (N, n=10) was detected by IHC using anti-SPHK2 and anti-hTERT antibodies. Statistical analyses of the data (shown in D) were performed using the Wilcoxon ran sum test, *P=0.0012, **P=0.031 (right panels).

Supplemental Figure 3. Inhibition or shRNA-mediated knockdown of SPHK2 induces telomere damage and caspase 3 activation in growth arrested A549 cells. (A-B) Effects of ABC294640 (ABC) (A) or shRNA-mediated knockdown of SPHK2 (B) on telomere damage (upper panels) using TIF assay, and caspase 3 activation using anti-caspase 3 antibody that detects activated caspase 3 (lower panels) in growth arrested A549 cells in response to serum starvation were measured by immunofluorescence. Data represent at least three independent experiments. (C-E) Cell cycle profiles of A549 cells in response to serum starvation (C-D) compared to controls, cells grown in complete media with serum (E), with/without ABC294640, ABC (C) or shRNA-SPHK2, SK2 (D) were measured using flow cytometry. Data represent three independent experiments, and error bars represent standard deviation.

Supplemental Figure 4. Effects of SPHK2 knockdown on ceramide, dihydro-ceramide and S1P were measured by lipidomics. (A-B) Effects of shRNA-mediated knockdown of SPHK2, SK2 mRNA, measured by qPCR (A) on sphingolipids (n=4, *P=0.02), such as ceramide, dihydro-C16-ceramide, sphingosine, dihydro-sphingosine, dihydro-S1P, or S1P were measured using mass spectrometry-based lipidomics (B) in serum starved and G0/G1 arrested A549 cells (n=6). Data are means ± SD (std dev) analyzed by student’s t-test.

Supplemental Figure 5. ABC294640 induces telomere damage but not caspase 3 activation in primary non-cancerous human lung fibroblasts. Effects of ABC294640 on telomere damage (upper panel) and caspase 3 activation (lower panel) in human non-cancerous lung epithelial cells were measured using TIF assay and anti-caspase 3 antibody that recognizes activated caspase 3, respectively, were measured by immunofluorescence. Data represent at least three independent experiments.

Supplemental Figure 6. Targeting SPHK2 induces TCF21-mediated apoptosis and lung tumor suppression. (A) Effects of ABC294640-mediated SPHK2 inhibition in the abundance of genes involved in lung tumor growth or suppression in tissues obtained from SCID mice treated with ABC294640 compared to vehicle were measured by qPCR using Superarray. (B-C) Increase in TCF21 mRNA in response to SPHK2 (B) (n=3, *P=0.016) or hTERT (n=3, *P=0.024) (C) knockdown using shRNAs in A549 cells was measured by qPCR. ShRNA-mediated knockdown of SPHK2 or hTERT (lower panels in B and C, respectively) were confirmed by qPCR in A549 cells compared to Scr-shRNA-transfected controls. Data are means ± SD from three independent experiments, analyzed by student’s t-test (n=3, *P=0.038). (D) Effects of ectopic expression of WT-hTERT or hTERTD684A on TCF21 mRNA in response to shRNA-mediated knockdown of SPHK2 compared to vector-and/or Scr-shRNA-transfected control A549 cells were measured by qPCR. Data are means ± SD from three independent experiments, analyzed by student’s t-test. (n=3, *P=0.0481, **P=0.0174, ***P=0.0267) (E) Knockdown of TCF21 using shRNA was confirmed in A549 cells in the presence of ABC294640 compared to Scr-transfected controls by qPCR. Data are means ± SD from three independent experiments, analyzed by student’s t-test (n=3, *P=0.029, **P=0.035). (F) Effects of shRNA-mediated knockdown of TCF21 on caspase 3 activation in the absence/presence of ABC294640 were measured by immunofluorescence using anti-caspase 3 antibody that detects activated caspase 3 in A549 cells. Images were quantified using ImageJ (bottom panel). Data

2 are means ± SD from three independent experiments, analyzed by student’s t-test (n=3, *P=0.022, **P=0.036, ***P=0.019). Scale bars represent 100 µm.

Supplemental Figure 7. Recombinant purified p16 and caspase 3 proteins interact with each other in vitro. (A) Association between recombinant purified p16 with caspase 3 (A) or mutant (pro) caspase 3 (B) was measured using SPR. (B) Purified caspase 3 protein is visualized by SDS-PAGE and Coomassie blue staining.